classification-optimization.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ca6a139d-c7ce-4bda-bf18-bc6242541452",
   "metadata": {},
   "source": [
    "# Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "6916b743-614c-458c-8bbe-0e0b3dc4265f",
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "53c8bd6c-c878-4ba3-99a3-274caa76e703",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "057345cef457482e8efabfae67324765",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(HTML(value='<center> <img\\nsrc=https://huggingface.co/front/assets/huggingface_logo-noborder.sv…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from huggingface_hub import login, whoami\n",
    "\n",
    "login()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "24205295-ad1e-4e33-9f7e-87f872d666a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# DATASET CONFIG\n",
    "DATASET='tyqiangz/multilingual-sentiments'\n",
    "MODEL='lxyuan/distilbert-base-multilingual-cased-sentiments-student'\n",
    "MIN_TOKENS=50\n",
    "MAX_TOKENS=512\n",
    "SAMPLES=10_000\n",
    "MAX_VUS = 1024\n",
    "\n",
    "\n",
    "TEXT_COLUMN='text'\n",
    "DATASET_PATH='data/classification-dataset.json'\n",
    "K6_BIN = \"/usr/bin/k6\"\n",
    "\n",
    "HUB_DATASET_PATH = f'{whoami()[\"name\"]}/classification-ie-optimization'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5af6946-b3ed-42b1-9580-be1ecf13d956",
   "metadata": {},
   "source": [
    "# Data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "958c8215-ad18-4fb5-b786-b629667bbbb9",
   "metadata": {},
   "source": [
    "## Dataset Creation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "8d29c264-21ad-4e69-9c1a-585cf867c297",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.\n",
      "\u001b[32m2025-01-31 08:19:50.344\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36msrc.process_dataset\u001b[0m:\u001b[36msample_dataset\u001b[0m:\u001b[36m39\u001b[0m - \u001b[32m\u001b[1mSampled dataset down to 10000 samples\u001b[0m\n",
      "\u001b[32m2025-01-31 08:19:50.494\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.process_dataset\u001b[0m:\u001b[36msave_dataset\u001b[0m:\u001b[36m61\u001b[0m - \u001b[1mSaved dataset to data/classification-dataset.json in JSON format\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "from datasets import load_dataset\n",
    "from transformers import AutoTokenizer\n",
    "from src.process_dataset import tokenize_and_filter, sample_dataset, save_dataset\n",
    "\n",
    "dataset = load_dataset(DATASET, 'all', split='train')\n",
    "\n",
    "# Map numbers to text labels\n",
    "label_mapping = {0: 'positive', 1: 'neutral', 2: 'negative'}\n",
    "dataset = dataset.map(lambda example: {'label_text': label_mapping[example['label']]})\n",
    "\n",
    "tokenizer = AutoTokenizer.from_pretrained(MODEL)\n",
    "\n",
    "dataset = tokenize_and_filter(dataset, tokenizer, min_tokens=MIN_TOKENS, max_tokens=MAX_TOKENS, text_column=TEXT_COLUMN, num_proc=8)\n",
    "dataset = sample_dataset(dataset, n_samples=SAMPLES ,seed=42)\n",
    "save_dataset(dataset.select_columns([TEXT_COLUMN, 'label_text']), DATASET_PATH)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "95626755-e1b5-44f4-9be2-827fdd53da5c",
   "metadata": {},
   "source": [
    "## Dataset Exploration\n",
    "Here we can see what our data looks like. Its a rather small dataset, it might be interesting in the future to look at longer datasets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "269df143-1d98-46a2-9ae6-87ba27198340",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Min:\t50\n",
      "Max:\t509\n",
      "Mean:\t99.3866\n",
      "StdDev:\t59.00211238054564\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tokens = dataset.to_pandas().num_tokens\n",
    "tokens.hist(bins=100)\n",
    "print(f'Min:\\t{tokens.min()}')\n",
    "print(f'Max:\\t{tokens.max()}')\n",
    "print(f'Mean:\\t{tokens.mean()}')\n",
    "print(f'StdDev:\\t{tokens.std()}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "199632b1-8a67-4c7c-9573-7465c011cbf4",
   "metadata": {},
   "source": [
    "# Experiments"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0f866bd3-430c-41ca-839f-7c7f7a546b16",
   "metadata": {},
   "source": [
    "## Experiment Structure\n",
    "The config structure should be read like this:\n",
    "\n",
    "```mermaid\n",
    "graph TD;\n",
    "    A[\"🖥️ InstanceConfig <br> Defines hardware, scaling, and GPUs\"]:::instance\n",
    "    A -->|Contains| B[\"🛳️ ImageConfig <br> Specifies Docker container and model URL\"]:::image\n",
    "    B -->|Contains| C[\"⚙️InfinityConfig <br> Configures Infinity settings like batch size\"]:::infinity\n",
    "\n",
    "    %% Define consistent styles with improved readability\n",
    "    classDef instance fill:#f9c74f,stroke:#333,stroke-width:2px;\n",
    "    classDef image fill:#90be6d,stroke:#333,stroke-width:2px;\n",
    "    classDef infinity fill:#4A90E2,stroke:#333,stroke-width:2px;\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "63763da5-eaef-4ce8-a7a9-83ffdb6911d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "from dataclasses import dataclass, field, asdict\n",
    "from typing import Dict\n",
    "\n",
    "@dataclass\n",
    "class InstanceConfig:\n",
    "    repository: str\n",
    "    accelerator: str\n",
    "    instance_size: str\n",
    "    instance_type: str\n",
    "    custom_image: Dict[str, str]  # Dict of str -> str\n",
    "    vendor: str\n",
    "    region: str\n",
    "\n",
    "@dataclass\n",
    "class InfinityConfig:\n",
    "    INFINITY_PORT: str = \"80\"\n",
    "    INFINITY_ENGINE: str = \"torch\"\n",
    "    INFINITY_BATCH_SIZE: str = \"16\"\n",
    "    INFINITY_DTYPE: str = \"auto\"\n",
    "    INFINITY_EMBEDDING_DTYPE: str = \"float32\"\n",
    "    INFINITY_POOLING_METHOD: str = \"auto\"\n",
    "    INFINITY_COMPILE: str = 'true'\n",
    "    INFINITY_BETTERTRANSFORMER: str = 'true'\n",
    "    INFINITY_MODEL_ID: str = \"/repository\"\n",
    "\n",
    "    def to_dict(self) -> Dict[str, str]:\n",
    "        \"\"\"Converts the dataclass to a dictionary representation.\"\"\"\n",
    "        return asdict(self)\n",
    "\n",
    "@dataclass\n",
    "class ImageConfig:\n",
    "    health_route: str = \"/health\"\n",
    "    url: str = \"michaelf34/infinity:0.0.75-trt-onnx\"\n",
    "    env: InfinityConfig = field(default_factory=InfinityConfig)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "cf2734e8-4ac9-41cc-9654-e2bfe7b5f19d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def set_env(batch_size: int, image_config: ImageConfig) -> Dict[str, str]:\n",
    "    config = asdict(image_config)\n",
    "    config['env'][\"INFINITY_BATCH_SIZE\"] = str(batch_size)\n",
    "    return config"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "97f02d7d-9bbc-44ca-aefe-3e66f302ea18",
   "metadata": {},
   "source": [
    "`deploy_endpoint` gives some basic features, you might want to expand this functionality for your experiments though!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "1ae77ff1-9e01-4837-8035-5d39257ef5d1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\u001b[0;31mSignature:\u001b[0m \u001b[0mdeploy_endpoint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minstance_config\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mendpoint_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwait\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
       "\u001b[0;31mDocstring:\u001b[0m Creates and deploys an inference endpoint using the given instance configuration.\n",
       "\u001b[0;31mFile:\u001b[0m      ~/encoder-analysis/src/deployment.py\n",
       "\u001b[0;31mType:\u001b[0m      function"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from src.deployment import deploy_endpoint\n",
    "?deploy_endpoint"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cf3743a9-8903-4792-b68b-f77c816070a8",
   "metadata": {},
   "source": [
    "## Run Experiments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "6ec56d04-0eb0-4065-ac86-d310a00ee2b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "instance_config_experiment_dicts = [\n",
    "    {\n",
    "        'accelerator': 'gpu',\n",
    "        'instance_size': 'x1',\n",
    "        'instance_type': 'nvidia-l4',\n",
    "        'vendor': 'aws',\n",
    "        'region': 'us-east-1',\n",
    "    },\n",
    "    {\n",
    "        'accelerator': 'gpu',\n",
    "        'instance_size': 'x1',\n",
    "        'instance_type': 'nvidia-t4',\n",
    "        'vendor': 'aws',\n",
    "        'region': 'us-east-1',\n",
    "    },\n",
    "]\n",
    "instance_config_experiments = [InstanceConfig(repository=MODEL, **instance_config_experiment_dict, custom_image=None) for instance_config_experiment_dict in instance_config_experiment_dicts]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "64e8f4d4-cc7b-4414-9183-d84173ae1bf6",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[32m2025-01-30 13:47:00.302\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=32, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:00.708\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:01.240\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:01.241\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=64, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:01.673\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:01.777\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:01.778\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=128, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:02.140\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:02.332\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:02.332\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=256, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:02.592\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:02.724\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:02.724\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=512, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:02.874\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:03.021\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:03.022\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=1024, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:03.201\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:03.473\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:33.689\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:33.690\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:33.690\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:33.690\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:33.691\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:33.691\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:50.870\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:50.871\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 603.16 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:47:50.871\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:05.482\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:05.483\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 722.80 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:05.483\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:19.300\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_128.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:19.301\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 793.22 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:19.301\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:33.536\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_256.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:33.537\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 823.11 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:33.537\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:49.556\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_512.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:49.557\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 512 VUs: 812.27 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:49.557\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:49.557\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 13:48:49.558\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 384\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:04.636\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_384.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:04.637\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 384 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:04.637\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 319\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:19.514\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_319.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:19.515\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 319 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:19.515\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 287\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:33.938\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_287.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:33.939\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 287 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:33.939\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 271\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:48.114\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_271.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:48.115\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 832.28 req/sec with 271 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:49:48.116\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 279\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:02.670\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_279.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:02.671\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 279 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:02.672\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 275\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:16.973\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_275.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:16.974\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 275 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:16.975\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 273\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:31.405\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_32_273.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:31.406\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 273 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:31.406\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 272\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.829\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.829\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.830\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.830\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.830\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.831\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.834\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.834\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 599.97 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.835\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.838\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.838\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 707.45 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.839\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.841\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_128.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.842\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 766.11 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.842\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.845\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_256.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.845\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 847.60 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.846\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.848\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_512.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.849\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 512 VUs: 832.84 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.849\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.849\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.850\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 384\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.853\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_384.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.853\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 858.38 req/sec with 384 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.854\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 448\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.856\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_448.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.857\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 871.23 req/sec with 448 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.857\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 480\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.860\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_480.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.860\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 480 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.861\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 464\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.863\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_464.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.864\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 464 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.864\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 456\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.867\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_456.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.867\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 456 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.868\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 452\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.870\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_452.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.871\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 452 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.871\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 450\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.874\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_64_450.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.874\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 450 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:43.875\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 449\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.419\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.419\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.420\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.420\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.420\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.421\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.424\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.424\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 599.58 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.425\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.428\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.428\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 707.53 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.429\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.431\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_128.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.432\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 792.21 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.432\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.435\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_256.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.435\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 803.76 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.436\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.436\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.436\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 192\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.439\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_192.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.440\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 192 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.440\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 159\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.443\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_159.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.443\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 159 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.444\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 143\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.446\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_143.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.447\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 143 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.447\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 135\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.450\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_135.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.450\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 135 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.450\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 131\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.453\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_131.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.454\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 131 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.454\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 129\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.457\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_128_129.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.458\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 129 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:50:54.458\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.146\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.147\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.147\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.148\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.148\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.148\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.151\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.152\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 588.41 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.152\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.155\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.156\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 695.14 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.156\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.159\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_128.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.159\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 769.36 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.159\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.162\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_256.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.163\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 786.29 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.163\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.166\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_512.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.166\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 512 VUs: 827.81 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.167\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 1024\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.169\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_1024.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.170\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 1024 VUs: 776.52 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.170\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.171\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.171\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 768\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.174\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_768.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.174\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 768 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.175\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 639\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.177\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_639.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.178\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 639 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.178\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 575\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.181\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_575.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.181\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 575 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.182\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 543\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.184\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_543.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.185\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 543 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.185\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 527\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.188\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_527.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.189\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 839.74 req/sec with 527 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.189\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 535\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.192\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_535.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.192\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 535 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.192\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 531\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.195\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_531.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.196\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 531 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.196\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 529\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.199\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_256_529.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.199\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 529 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:06.199\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 528\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.245\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.246\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.246\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.246\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.247\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.247\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.250\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.251\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 582.03 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.251\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.254\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.254\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 705.80 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.255\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.258\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_128.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.258\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 783.31 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.258\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.261\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_256.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.262\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 794.71 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.262\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.262\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.263\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 192\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.265\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_192.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.266\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 805.47 req/sec with 192 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.266\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 224\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.269\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_224.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.269\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 224 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.270\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 208\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.273\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_208.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.273\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 208 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.274\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 200\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.276\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_200.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.277\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 200 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.277\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 196\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.280\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_196.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.280\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 818.73 req/sec with 196 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.281\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 198\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.283\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_512_198.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.284\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 198 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:17.284\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 197\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.553\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.554\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.554\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.555\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 1024\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.555\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.555\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.559\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.559\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 595.08 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.559\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.562\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.563\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 712.44 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.563\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.566\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_128.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.567\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 785.24 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.567\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.570\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_256.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.570\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 776.26 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.570\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.571\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.571\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 192\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.574\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_192.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.574\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 787.00 req/sec with 192 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.575\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 224\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.578\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_224.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.578\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 803.92 req/sec with 224 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.578\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 240\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.581\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_240.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.582\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 240 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.582\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 232\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.585\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_232.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.585\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 232 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.585\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 228\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.588\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_228.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.589\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 816.07 req/sec with 228 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.589\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 230\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.592\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_trt-onnx_torch_1024_230.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.592\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 230 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:27.592\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 229\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:37.635\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=32, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:37.872\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:37.985\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:37.986\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=64, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.184\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.302\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.303\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=128, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.398\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.543\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.544\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=256, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.629\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.697\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.698\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=512, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.846\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.973\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:38.973\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=1024, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:39.154\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:51:39.216\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:19.865\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:19.866\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:19.866\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:19.866\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:19.867\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:19.867\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:37.399\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:37.400\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 588.21 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:37.400\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:52.201\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:52.202\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 712.36 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:52:52.203\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:05.788\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_128.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:05.789\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 806.21 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:05.790\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:23.682\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_256.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:23.683\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 630.69 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:23.684\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:23.684\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:23.685\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 192\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:39.245\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_192.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:39.245\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 192 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:39.246\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 159\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:53.168\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_159.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:53.169\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 159 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:53:53.169\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 143\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:07.593\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_143.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:07.594\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 143 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:07.595\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 135\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:21.449\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_135.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:21.450\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 135 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:21.451\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 131\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:38.882\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_131.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:38.883\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 131 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:38.883\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 129\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:52.491\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_32_129.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:52.492\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 807.15 req/sec with 129 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:54:52.493\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 130\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:09.539\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:09.540\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:09.540\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:09.541\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:09.541\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:09.541\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:31.041\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:31.042\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 476.95 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:31.043\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:45.894\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:45.895\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 709.38 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:45.896\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:59.793\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_128.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:59.794\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 786.27 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:55:59.795\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:14.135\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_256.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:14.136\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 816.46 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:14.136\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:33.135\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_512.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:33.136\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 512 VUs: 653.63 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:33.136\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:33.136\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:33.137\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 384\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:47.802\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_384.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:47.803\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 846.20 req/sec with 384 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:56:47.804\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 448\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:02.641\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_448.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:02.642\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 875.50 req/sec with 448 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:02.642\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 480\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:18.007\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_480.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:18.008\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 480 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:18.008\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 464\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:37.157\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_464.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:37.158\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 464 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:37.159\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 456\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:52.352\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_456.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:52.353\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 456 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:57:52.353\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 452\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:07.382\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_452.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:07.383\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 452 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:07.384\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 450\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:22.512\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_64_450.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:22.513\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 450 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:22.514\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 449\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:44.150\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:44.151\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:44.151\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:44.152\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:44.152\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 13:58:44.152\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 13:59:08.490\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:59:08.491\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 419.97 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:59:08.492\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 13:59:23.827\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:59:23.828\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 686.24 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:59:23.828\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 13:59:38.081\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_128.json\u001b[0m\n",
      "\u001b[32m2025-01-30 13:59:38.082\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 766.11 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 13:59:38.083\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:02.113\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_256.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:02.114\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 455.13 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:02.115\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:02.115\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:02.115\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 192\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:21.715\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_192.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:21.716\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 192 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:21.716\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 159\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:36.287\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_159.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:36.287\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 159 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:36.288\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 143\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:51.023\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_143.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:51.024\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 143 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:00:51.025\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 135\u001b[0m\n",
      "\u001b[32m2025-01-30 14:01:15.242\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_135.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:01:15.243\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 135 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:01:15.244\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 131\u001b[0m\n",
      "\u001b[32m2025-01-30 14:01:36.792\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_131.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:01:36.793\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 131 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:01:36.794\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 129\u001b[0m\n",
      "\u001b[32m2025-01-30 14:01:51.824\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_128_129.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:01:51.825\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 129 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:01:51.825\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 14:02:19.867\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:02:19.868\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:02:19.868\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:02:19.868\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 14:02:19.869\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:02:19.869\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:02:49.375\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:02:49.375\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 345.14 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:02:49.376\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:06.258\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:06.259\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 620.62 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:06.259\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:22.080\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_128.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:22.081\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 682.73 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:22.081\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:49.035\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_256.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:49.036\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 402.22 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:49.037\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:49.037\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:03:49.037\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 192\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:05.695\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_192.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:05.696\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 192 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:05.696\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 159\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:21.800\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_159.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:21.801\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 159 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:21.801\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 143\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:37.778\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_143.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:37.780\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 143 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:37.780\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 135\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:57.561\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_135.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:57.562\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 135 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:04:57.562\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 131\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:14.040\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_131.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:14.041\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 131 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:14.041\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 129\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:30.015\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_256_129.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:30.016\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 129 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:30.016\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:40.529\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:40.529\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:40.530\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:40.530\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:40.530\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:40.531\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:57.560\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:57.562\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 605.87 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:05:57.562\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:12.161\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:12.162\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 722.35 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:12.163\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:26.338\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_128.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:26.339\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 774.31 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:26.340\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:40.705\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_256.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:40.706\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 813.92 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:40.706\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:59.225\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_512.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:59.226\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 512 VUs: 673.87 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:59.226\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:59.227\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:06:59.227\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 384\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:15.269\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_384.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:15.270\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 384 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:15.271\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 319\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:30.374\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_319.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:30.375\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 319 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:30.375\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 287\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:44.761\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_287.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:44.762\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 826.63 req/sec with 287 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:44.762\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 303\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:59.745\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_303.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:59.746\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 303 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:07:59.746\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 295\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:14.776\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_295.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:14.777\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 295 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:14.777\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 291\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:29.696\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_291.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:29.697\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 291 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:29.697\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 289\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:44.638\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_512_289.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:44.639\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 289 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:44.639\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 288\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:55.039\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:55.039\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:55.039\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-l4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:55.040\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 1024\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:55.040\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:08:55.040\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:09:13.096\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:09:13.097\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 570.66 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:09:13.098\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:09:29.539\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:09:29.539\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 637.29 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:09:29.540\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 14:09:45.336\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_128.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:09:45.337\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 128 VUs: 682.80 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:09:45.337\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:01.283\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_256.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:01.284\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 256 VUs: 720.01 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:01.284\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:17.740\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_512.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:17.741\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 512 VUs: 783.24 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:17.741\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 1024\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:38.560\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_1024.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:38.561\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 1024 VUs: 730.86 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:38.562\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:38.562\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:38.562\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 768\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:57.642\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_768.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:57.643\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 768 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:10:57.643\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 639\u001b[0m\n",
      "\u001b[32m2025-01-30 14:11:15.212\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_639.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:11:15.213\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 639 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:11:15.214\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 575\u001b[0m\n",
      "\u001b[32m2025-01-30 14:11:32.718\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_575.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:11:32.720\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 575 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:11:32.721\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 543\u001b[0m\n",
      "\u001b[32m2025-01-30 14:11:50.468\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_543.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:11:50.469\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 543 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:11:50.470\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 527\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:06.998\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_527.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:06.999\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 787.44 req/sec with 527 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:06.999\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 535\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:24.084\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_535.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:24.085\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 535 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:24.086\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 531\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:41.267\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_531.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:41.268\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 531 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:41.269\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 529\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:58.620\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-l4/aws_nvidia-l4_default_torch_1024_529.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:58.621\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 529 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:12:58.622\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 528\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:08.680\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=32, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:08.803\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:08.889\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:08.889\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=64, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.004\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.087\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.088\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=128, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.170\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.287\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.288\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=256, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.369\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.434\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.434\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=512, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.514\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.585\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.586\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=1024, image_url=michaelf34/infinity:0.0.75-trt-onnx\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.778\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:09.855\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.075\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.076\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.076\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.077\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.077\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.077\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.080\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_32_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.081\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 218.64 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.081\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.084\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_32_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.084\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 206.25 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.085\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.085\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.085\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.088\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_32_48.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.088\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.089\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 39\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.091\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_32_39.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.092\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 39 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.093\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 35\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.095\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_32_35.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.096\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 35 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.096\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 33\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.099\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_32_33.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.099\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 33 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:13:50.100\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.181\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.182\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.182\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.183\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.183\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.183\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.186\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_64_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.187\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 222.68 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.187\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.190\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_64_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.191\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 216.53 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.191\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.191\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.192\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.194\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_64_48.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.195\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.195\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 39\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.198\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_64_39.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.199\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 39 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.199\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 35\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.202\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_64_35.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.202\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 35 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.203\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 33\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.205\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_64_33.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.206\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 33 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:00.206\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.302\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.302\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.303\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.303\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.303\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.304\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.307\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_128_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.308\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 216.21 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.308\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.310\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_128_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.311\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 216.41 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.311\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.312\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.312\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.315\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_128_48.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.315\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.316\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 39\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.318\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_128_39.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.319\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 39 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.319\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 35\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.322\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_128_35.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.322\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 35 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.323\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 33\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.325\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_128_33.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.326\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 33 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:10.326\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.400\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.401\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.401\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.402\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.402\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.402\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.405\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_256_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.406\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 215.76 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.406\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.409\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_256_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.409\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 217.74 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.410\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.410\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.410\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.413\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_256_48.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.414\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 223.66 req/sec with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.414\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 56\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.417\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_256_56.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.417\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 56 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.418\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 52\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.420\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_256_52.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.421\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 52 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.421\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 50\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.424\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_256_50.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.424\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 50 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:20.425\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 49\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.545\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.546\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.546\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.546\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.547\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.547\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.550\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_512_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.551\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 224.41 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.551\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.554\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_512_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.554\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 224.10 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.555\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.555\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.555\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.558\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_512_48.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.558\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 227.29 req/sec with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.559\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 56\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.561\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_512_56.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.562\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 56 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.563\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 52\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.565\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_512_52.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.566\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 52 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.566\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 50\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.569\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_512_50.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.569\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 50 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:30.570\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 49\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.651\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.651\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.652\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.652\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 1024\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.652\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.653\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.656\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_1024_32.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.656\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 212.86 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.657\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.660\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_1024_64.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.660\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 210.25 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.660\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.661\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.661\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.664\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_1024_48.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.664\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.665\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 39\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.668\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_1024_39.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.668\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 39 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.668\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 35\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.671\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_1024_35.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.672\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 35 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.672\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 33\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.675\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m42\u001b[0m - \u001b[1mresults file /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_trt-onnx_torch_1024_33.json already exists\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.675\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 214.24 req/sec with 33 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:40.676\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 34\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:50.714\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=32, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:50.866\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:50.954\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:50.955\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=64, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.054\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.125\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.126\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=128, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.224\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.296\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.297\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=256, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.377\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.436\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.437\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=512, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.510\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.583\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.584\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m34\u001b[0m - \u001b[1mRunning experiment with batch_size=1024, image_url=michaelf34/infinity:0.0.75\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.723\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m21\u001b[0m - \u001b[1mCreating inference endpoint...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:14:51.814\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.deployment\u001b[0m:\u001b[36mdeploy_endpoint\u001b[0m:\u001b[36m40\u001b[0m - \u001b[1mWaiting for endpoint to be ready...\u001b[0m\n",
      "\u001b[32m2025-01-30 14:19:27.892\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:19:27.892\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:19:27.893\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:19:27.893\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:19:27.893\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:19:27.894\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:20:13.275\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_32_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:20:13.276\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 223.00 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:20:13.276\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:20:59.453\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_32_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:20:59.455\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 220.20 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:20:59.456\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:20:59.456\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:20:59.456\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:21:46.730\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_32_48.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:21:46.731\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:21:46.731\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 39\u001b[0m\n",
      "\u001b[32m2025-01-30 14:22:32.968\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_32_39.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:22:32.969\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 39 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:22:32.970\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 35\u001b[0m\n",
      "\u001b[32m2025-01-30 14:23:20.265\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_32_35.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:23:20.266\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 35 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:23:20.267\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 33\u001b[0m\n",
      "\u001b[32m2025-01-30 14:24:06.516\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_32_33.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:24:06.517\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 33 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:24:06.518\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:24:16.602\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:24:16.603\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:24:16.603\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:24:16.604\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:24:16.604\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:24:16.605\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:25:05.556\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_64_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:25:05.557\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 206.61 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:25:05.557\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:25:54.950\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_64_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:25:54.951\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 205.63 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:25:54.951\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:25:54.951\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:25:54.952\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:26:44.789\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_64_48.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:26:44.790\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:26:44.791\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 39\u001b[0m\n",
      "\u001b[32m2025-01-30 14:27:33.408\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_64_39.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:27:33.409\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 208.27 req/sec with 39 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:27:33.410\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 43\u001b[0m\n",
      "\u001b[32m2025-01-30 14:28:22.339\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_64_43.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:28:22.340\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 43 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:28:22.340\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 41\u001b[0m\n",
      "\u001b[32m2025-01-30 14:29:12.009\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_64_41.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:29:12.010\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 41 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:29:12.010\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 40\u001b[0m\n",
      "\u001b[32m2025-01-30 14:29:22.104\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:29:22.104\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:29:22.104\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:29:22.105\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 128\u001b[0m\n",
      "\u001b[32m2025-01-30 14:29:22.105\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:29:22.105\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:30:08.797\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_128_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:30:08.797\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 216.75 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:30:08.798\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:30:59.403\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_128_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:30:59.404\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 200.67 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:30:59.405\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:30:59.405\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:30:59.405\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:31:47.058\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_128_48.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:31:47.059\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:31:47.060\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 39\u001b[0m\n",
      "\u001b[32m2025-01-30 14:32:35.746\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_128_39.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:32:35.747\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 39 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:32:35.747\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 35\u001b[0m\n",
      "\u001b[32m2025-01-30 14:33:23.541\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_128_35.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:33:23.542\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 35 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:33:23.543\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 33\u001b[0m\n",
      "\u001b[32m2025-01-30 14:34:11.172\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_128_33.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:34:11.173\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 33 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:34:11.173\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:34:21.308\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:34:21.309\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:34:21.309\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:34:21.310\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 256\u001b[0m\n",
      "\u001b[32m2025-01-30 14:34:21.310\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:34:21.311\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:35:07.827\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_256_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:35:07.828\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 217.59 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:35:07.828\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:35:54.735\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_256_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:35:54.736\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 216.78 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:35:54.737\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:35:54.737\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:35:54.737\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:36:41.047\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_256_48.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:36:41.049\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 219.11 req/sec with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:36:41.049\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 56\u001b[0m\n",
      "\u001b[32m2025-01-30 14:37:30.433\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_256_56.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:37:30.433\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 56 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:37:30.434\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 52\u001b[0m\n",
      "\u001b[32m2025-01-30 14:38:18.272\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_256_52.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:38:18.273\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 52 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:38:18.273\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 50\u001b[0m\n",
      "\u001b[32m2025-01-30 14:39:06.215\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_256_50.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:39:06.216\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 50 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:39:06.216\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 49\u001b[0m\n",
      "\u001b[32m2025-01-30 14:39:16.407\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:39:16.407\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:39:16.408\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:39:16.408\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 512\u001b[0m\n",
      "\u001b[32m2025-01-30 14:39:16.408\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:39:16.408\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:40:02.743\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_512_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:40:02.744\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 218.37 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:40:02.745\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:40:49.726\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_512_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:40:49.727\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 216.48 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:40:49.727\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:40:49.728\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:40:49.728\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:41:36.321\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_512_48.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:41:36.322\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:41:36.322\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 39\u001b[0m\n",
      "\u001b[32m2025-01-30 14:42:22.140\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_512_39.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:42:22.141\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 221.22 req/sec with 39 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:42:22.141\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 43\u001b[0m\n",
      "\u001b[32m2025-01-30 14:43:08.828\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_512_43.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:43:08.828\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 43 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:43:08.829\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 41\u001b[0m\n",
      "\u001b[32m2025-01-30 14:43:54.372\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_512_41.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:43:54.373\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 222.59 req/sec with 41 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:43:54.373\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 42\u001b[0m\n",
      "\u001b[32m2025-01-30 14:44:04.734\u001b[0m | \u001b[32m\u001b[1mSUCCESS \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m47\u001b[0m - \u001b[32m\u001b[1mEndpoint is ready!\u001b[0m\n",
      "\u001b[32m2025-01-30 14:44:04.735\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m48\u001b[0m - \u001b[1m\tVendor: aws\u001b[0m\n",
      "\u001b[32m2025-01-30 14:44:04.735\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m49\u001b[0m - \u001b[1m\tInstance: nvidia-t4\u001b[0m\n",
      "\u001b[32m2025-01-30 14:44:04.736\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36m__main__\u001b[0m:\u001b[36m<module>\u001b[0m:\u001b[36m50\u001b[0m - \u001b[1m\tBatch Size: 1024\u001b[0m\n",
      "\u001b[32m2025-01-30 14:44:04.736\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m91\u001b[0m - \u001b[1mStarting exponential search for optimal VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:44:04.736\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 32\u001b[0m\n",
      "\u001b[32m2025-01-30 14:44:54.004\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_1024_32.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:44:54.005\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 32 VUs: 206.07 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:44:54.005\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m93\u001b[0m - \u001b[1mTesting with VUs: 64\u001b[0m\n",
      "\u001b[32m2025-01-30 14:45:43.647\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_1024_64.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:45:43.647\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m96\u001b[0m - \u001b[1mThroughput for 64 VUs: 204.61 req/sec\u001b[0m\n",
      "\u001b[32m2025-01-30 14:45:43.648\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m100\u001b[0m - \u001b[1mThroughput improvement is less than 2%, stopping exponential search.\u001b[0m\n",
      "\u001b[32m2025-01-30 14:45:43.648\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m108\u001b[0m - \u001b[1mStarting binary search refinement\u001b[0m\n",
      "\u001b[32m2025-01-30 14:45:43.649\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 48\u001b[0m\n",
      "\u001b[32m2025-01-30 14:46:31.936\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_1024_48.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:46:31.937\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m116\u001b[0m - \u001b[1mThroughput improved to 210.01 req/sec with 48 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:46:31.937\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 56\u001b[0m\n",
      "\u001b[32m2025-01-30 14:47:21.470\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_1024_56.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:47:21.471\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 56 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:47:21.472\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 52\u001b[0m\n",
      "\u001b[32m2025-01-30 14:48:11.327\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_1024_52.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:48:11.328\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 52 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:48:11.328\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m112\u001b[0m - \u001b[1mTesting with VUs: 50\u001b[0m\n",
      "\u001b[32m2025-01-30 14:49:00.939\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36mcall_k6\u001b[0m:\u001b[36m75\u001b[0m - \u001b[1mResults written to /home/ec2-user/encoder-analysis/results/classification/nvidia-t4/aws_nvidia-t4_default_torch_1024_50.json\u001b[0m\n",
      "\u001b[32m2025-01-30 14:49:00.940\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m120\u001b[0m - \u001b[1mThroughput did not improve with 50 VUs\u001b[0m\n",
      "\u001b[32m2025-01-30 14:49:00.940\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36msrc.k6\u001b[0m:\u001b[36moptimal_vus\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1mOptimal VUs: 49\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "import itertools\n",
    "import copy\n",
    "from pathlib import Path\n",
    "from time import sleep\n",
    "\n",
    "from loguru import logger\n",
    "\n",
    "from src.k6 import call_k6, optimal_vus, image_dict\n",
    "\n",
    "template_file = \"classification-analysis.js.j2\"\n",
    "output_file = Path(\"./generated\").resolve() / \"classification-analysis.js\"\n",
    "\n",
    "start_vus = 32\n",
    "batch_sizes = [16, 32, 64, 128, 256, 512, 1024]\n",
    "image_urls = ['michaelf34/infinity:0.0.75-trt-onnx', 'michaelf34/infinity:0.0.75']\n",
    "\n",
    "# Create the cross-product of batch sizes and image URLs\n",
    "experiments = list(itertools.product(batch_sizes, image_urls))\n",
    "\n",
    "for og_instance_config_experiment in instance_config_experiments:\n",
    "    for image_url in image_urls:\n",
    "        endpoints = []\n",
    "        for batch_size in batch_sizes:\n",
    "            # Configure Infinity Settings, pass them up to the image\n",
    "            infinty_config = InfinityConfig(INFINITY_BATCH_SIZE=str(batch_size))\n",
    "            image_config = ImageConfig(env=infinty_config, url=image_url)\n",
    "    \n",
    "            # Add the image to the HW instance\n",
    "            instance_config_experiment = copy.deepcopy(og_instance_config_experiment)\n",
    "            instance_config_experiment.custom_image = image_config\n",
    "    \n",
    "            # Deploy the endpoint\n",
    "            logger.info(f\"Running experiment with batch_size={batch_size}, image_url={image_url}\")\n",
    "            vendor = instance_config_experiment.vendor\n",
    "            image_url_short = image_dict[image_url]\n",
    "            endpoint = deploy_endpoint(instance_config_experiment, endpoint_name=f'ea-{vendor}-{image_url_short}-{batch_size}', wait=False)\n",
    "            endpoints.append(endpoint)\n",
    "    \n",
    "        endpoints[0].wait()\n",
    "        \n",
    "        for endpoint in endpoints:\n",
    "            endpoint.fetch()\n",
    "            batch_size = endpoint.__dict__[\"raw\"][\"model\"][\"env\"][\"INFINITY_BATCH_SIZE\"]\n",
    "            vendor = endpoint.__dict__[\"raw\"][\"provider\"][\"vendor\"]\n",
    "            instance_type = endpoint.__dict__[\"raw\"][\"compute\"][\"instanceType\"]\n",
    "            logger.success(f'Endpoint is ready!')\n",
    "            logger.info(f'\\tVendor: {vendor}')\n",
    "            logger.info(f'\\tInstance: {instance_type}')\n",
    "            logger.info(f'\\tBatch Size: {batch_size}')\n",
    "    \n",
    "            args_dict = dict(\n",
    "                text_column=TEXT_COLUMN,\n",
    "                endpoint=endpoint,\n",
    "                total_requests=SAMPLES,\n",
    "                template_file=template_file,\n",
    "                output_file=output_file,\n",
    "                dataset_path=DATASET_PATH,\n",
    "                k6_bin=K6_BIN\n",
    "            )\n",
    "    \n",
    "            optimal_vus(max_vus=MAX_VUS, args_dict=args_dict, start_vus=start_vus)\n",
    "            sleep(5)\n",
    "            endpoint.delete()\n",
    "            sleep(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee3a92a9-f53c-4361-a468-7ed83853be9c",
   "metadata": {},
   "source": [
    "## Process Experiments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "ab1162f7-bb4a-49a0-9dfe-7390f92df39b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>total_requests</th>\n",
       "      <th>test_duration_sec</th>\n",
       "      <th>successful_requests</th>\n",
       "      <th>avg_latency_ms</th>\n",
       "      <th>p95_latency_ms</th>\n",
       "      <th>accuracy_percentage</th>\n",
       "      <th>throughput_req_per_sec</th>\n",
       "      <th>hw_type</th>\n",
       "      <th>batch_size</th>\n",
       "      <th>image</th>\n",
       "      <th>engine</th>\n",
       "      <th>vendor</th>\n",
       "      <th>vus</th>\n",
       "      <th>seconds_per_billion</th>\n",
       "      <th>cost_per_sec</th>\n",
       "      <th>1B_cost</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>194</th>\n",
       "      <td>10000</td>\n",
       "      <td>11.422015</td>\n",
       "      <td>10000</td>\n",
       "      <td>502.9814</td>\n",
       "      <td>743.0</td>\n",
       "      <td>58.03</td>\n",
       "      <td>875.502269</td>\n",
       "      <td>nvidia-l4</td>\n",
       "      <td>64</td>\n",
       "      <td>michaelf34/infinity:0.0.75</td>\n",
       "      <td>torch</td>\n",
       "      <td>aws</td>\n",
       "      <td>448</td>\n",
       "      <td>1.142201e+06</td>\n",
       "      <td>0.000222</td>\n",
       "      <td>253.822554</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>106</th>\n",
       "      <td>10000</td>\n",
       "      <td>11.477969</td>\n",
       "      <td>10000</td>\n",
       "      <td>501.8690</td>\n",
       "      <td>740.0</td>\n",
       "      <td>58.01</td>\n",
       "      <td>871.234286</td>\n",
       "      <td>nvidia-l4</td>\n",
       "      <td>64</td>\n",
       "      <td>michaelf34/infinity:0.0.75-trt-onnx</td>\n",
       "      <td>torch</td>\n",
       "      <td>aws</td>\n",
       "      <td>448</td>\n",
       "      <td>1.147797e+06</td>\n",
       "      <td>0.000222</td>\n",
       "      <td>255.065974</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>107</th>\n",
       "      <td>10000</td>\n",
       "      <td>11.517683</td>\n",
       "      <td>10000</td>\n",
       "      <td>541.2980</td>\n",
       "      <td>779.0</td>\n",
       "      <td>58.03</td>\n",
       "      <td>868.230160</td>\n",
       "      <td>nvidia-l4</td>\n",
       "      <td>64</td>\n",
       "      <td>michaelf34/infinity:0.0.75-trt-onnx</td>\n",
       "      <td>torch</td>\n",
       "      <td>aws</td>\n",
       "      <td>480</td>\n",
       "      <td>1.151768e+06</td>\n",
       "      <td>0.000222</td>\n",
       "      <td>255.948517</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108</th>\n",
       "      <td>10000</td>\n",
       "      <td>11.542635</td>\n",
       "      <td>10000</td>\n",
       "      <td>525.3836</td>\n",
       "      <td>778.0</td>\n",
       "      <td>58.01</td>\n",
       "      <td>866.353341</td>\n",
       "      <td>nvidia-l4</td>\n",
       "      <td>64</td>\n",
       "      <td>michaelf34/infinity:0.0.75-trt-onnx</td>\n",
       "      <td>torch</td>\n",
       "      <td>aws</td>\n",
       "      <td>464</td>\n",
       "      <td>1.154263e+06</td>\n",
       "      <td>0.000222</td>\n",
       "      <td>256.502990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>105</th>\n",
       "      <td>10000</td>\n",
       "      <td>11.649862</td>\n",
       "      <td>10000</td>\n",
       "      <td>438.9456</td>\n",
       "      <td>641.0</td>\n",
       "      <td>58.02</td>\n",
       "      <td>858.379273</td>\n",
       "      <td>nvidia-l4</td>\n",
       "      <td>64</td>\n",
       "      <td>michaelf34/infinity:0.0.75-trt-onnx</td>\n",
       "      <td>torch</td>\n",
       "      <td>aws</td>\n",
       "      <td>384</td>\n",
       "      <td>1.164986e+06</td>\n",
       "      <td>0.000222</td>\n",
       "      <td>258.885820</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     total_requests  test_duration_sec  successful_requests  avg_latency_ms  \\\n",
       "194           10000          11.422015                10000        502.9814   \n",
       "106           10000          11.477969                10000        501.8690   \n",
       "107           10000          11.517683                10000        541.2980   \n",
       "108           10000          11.542635                10000        525.3836   \n",
       "105           10000          11.649862                10000        438.9456   \n",
       "\n",
       "     p95_latency_ms  accuracy_percentage  throughput_req_per_sec    hw_type  \\\n",
       "194           743.0                58.03              875.502269  nvidia-l4   \n",
       "106           740.0                58.01              871.234286  nvidia-l4   \n",
       "107           779.0                58.03              868.230160  nvidia-l4   \n",
       "108           778.0                58.01              866.353341  nvidia-l4   \n",
       "105           641.0                58.02              858.379273  nvidia-l4   \n",
       "\n",
       "     batch_size                                image engine vendor  vus  \\\n",
       "194          64           michaelf34/infinity:0.0.75  torch    aws  448   \n",
       "106          64  michaelf34/infinity:0.0.75-trt-onnx  torch    aws  448   \n",
       "107          64  michaelf34/infinity:0.0.75-trt-onnx  torch    aws  480   \n",
       "108          64  michaelf34/infinity:0.0.75-trt-onnx  torch    aws  464   \n",
       "105          64  michaelf34/infinity:0.0.75-trt-onnx  torch    aws  384   \n",
       "\n",
       "     seconds_per_billion  cost_per_sec     1B_cost  \n",
       "194         1.142201e+06      0.000222  253.822554  \n",
       "106         1.147797e+06      0.000222  255.065974  \n",
       "107         1.151768e+06      0.000222  255.948517  \n",
       "108         1.154263e+06      0.000222  256.502990  \n",
       "105         1.164986e+06      0.000222  258.885820  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from src.process_dataset import load_json_files\n",
    "\n",
    "# Define cost per hour considering both hardware type and vendor\n",
    "cost_per_hour = {\n",
    "    ('gcp', 'nvidia-t4'): 0.5,\n",
    "    ('gcp', 'nvidia-l4'): 0.7,\n",
    "    ('aws', 'nvidia-t4'): 0.5,\n",
    "    ('aws', 'nvidia-l4'): 0.8,\n",
    "}\n",
    "\n",
    "# Load dataset\n",
    "df = load_json_files('results/classification')\n",
    "\n",
    "# Compute cost\n",
    "df['seconds_per_billion'] = 1e9 / df['throughput_req_per_sec']\n",
    "df['cost_per_sec'] = df.apply(lambda row: cost_per_hour.get((row['vendor'], row['hw_type']), 0) / 3600, axis=1)\n",
    "df['1B_cost'] = df['seconds_per_billion'] * df['cost_per_sec']\n",
    "\n",
    "df.sort_values(by='1B_cost').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "61d5f882-d2ac-454f-99f1-cbfa4ca3afd9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5d490ac10457432fb1ba797ad5a7533a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Uploading the dataset shards:   0%|          | 0/1 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "87a6fe50deeb431bb52d256374210ae1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Creating parquet from Arrow format:   0%|          | 0/1 [00:00<?, ?ba/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2e4a4147445d41c4b903b7b232702c04",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "README.md:   0%|          | 0.00/862 [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "CommitInfo(commit_url='https://huggingface.co/datasets/derek-thomas/classification-ie-optimization/commit/a297e75d5b7ec8bf858a36ee43ee11083c2b5cbd', commit_message='Upload dataset', commit_description='', oid='a297e75d5b7ec8bf858a36ee43ee11083c2b5cbd', pr_url=None, repo_url=RepoUrl('https://huggingface.co/datasets/derek-thomas/classification-ie-optimization', endpoint='https://huggingface.co', repo_type='dataset', repo_id='derek-thomas/classification-ie-optimization'), pr_revision=None, pr_num=None)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from datasets import Dataset\n",
    "# Push to the hub\n",
    "dataset = Dataset.from_pandas(df)\n",
    "dataset.push_to_hub(HUB_DATASET_PATH)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea1b6c96-f637-4dec-b1aa-54b45be09862",
   "metadata": {},
   "source": [
    "## CPU Experiments\n",
    "This is left to the user to play with, I experimented early, but moved on since it was too expensive."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "36b042d5-ea78-452d-b99e-beee3f2b1521",
   "metadata": {},
   "outputs": [],
   "source": [
    "instance_config_experiment_dicts = [\n",
    "    {\n",
    "        'accelerator': 'cpu',\n",
    "        'instance_size': 'x1',\n",
    "        'instance_type': 'intel-spr'\n",
    "    }\n",
    "]\n",
    "\n",
    "instance_config_experiments = [InstanceConfig(repository=MODEL, **ic_exp_dict, custom_image=None) for ic_exp_dict in instance_config_experiment_dicts]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8f5e4e1f-c775-4229-9fbb-51782290cb1a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from itertools import product\n",
    "\n",
    "batch_sizes = [1, 2, 4, 8, 16, 32, 64]\n",
    "infinity_engine = ['optimum', 'torch']\n",
    "start_vus = 1\n",
    "\n",
    "cross_product = list(product(batch_sizes, infinity_engine))\n",
    "\n",
    "for og_instance_config_experiment in instance_config_experiments:\n",
    "    for batch_size, infinity_engine in cross_product:\n",
    "        # Configure Infinity Settings, pass them up to the image\n",
    "        infinty_config = InfinityConfig(INFINITY_BATCH_SIZE=str(batch_size))\n",
    "        image_config = ImageConfig(env=infinty_config)\n",
    "\n",
    "        # Add the image to the HW instance\n",
    "        instance_config_experiment = copy.deepcopy(og_instance_config_experiment)\n",
    "        instance_config_experiment.custom_image = image_config\n",
    "\n",
    "        # Deploy the endpoint\n",
    "        logger.info(f'Deploying with Batch Size:\\t{batch_size}')\n",
    "        logger.info(f'Deploying with Infinity Engine:\\t{infinity_engine}')\n",
    "        endpoint = deploy_endpoint(instance_config_experiment)\n",
    "\n",
    "        args_dict = dict(\n",
    "            endpoint=endpoint,\n",
    "            total_requests=10_000,\n",
    "            template_file=template_file,\n",
    "            output_file=output_file,\n",
    "            dataset_path=DATASET_PATH,\n",
    "            k6_bin=K6_BIN\n",
    "        )\n",
    "        \n",
    "        optimal_vus(max_vus=CPU_MAX_VUS, args_dict=args_dict, start_vus=start_vus)\n",
    "        endpoint.delete()\n",
    "        sleep(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b86958f3-623e-4b7c-95f2-b5f47ad0b2ff",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}