Provider updates and optimizations across multiple modules #2220

kqlio67 · 2024-09-11T14:24:33Z

New providers added

g4f/Provider/Prodia.py (Provider updates and optimizations across multiple modules #2220 (comment))
Provider with support for models for generating images
3Guofeng3_v34.safetensors [50f420de], absolutereality_V16.safetensors [37db0fc3], absolutereality_v181.safetensors [3d9d4d2b], amIReal_V41.safetensors [0a8a2e61], analog-diffusion-1.0.ckpt [9ca13f02], aniverse_v30.safetensors [579e6f85], anythingv3_0-pruned.ckpt [2700c435], anything-v4.5-pruned.ckpt [65745d25], anythingV5_PrtRE.safetensors [893e49b9], AOM3A3_orangemixs.safetensors [9600da17], blazing_drive_v10g.safetensors [ca1c1eab], breakdomain_I2428.safetensors [43cc7d2f], breakdomain_M2150.safetensors [15f7afca], cetusMix_Version35.safetensors [de2f2560], childrensStories_v13D.safetensors [9dfaabcb], childrensStories_v1SemiReal.safetensors [a1c56dbb], childrensStories_v1ToonAnime.safetensors [2ec7b88b], Counterfeit_v30.safetensors [9e2a8f19], cuteyukimixAdorable_midchapter3.safetensors [04bdffe6], cyberrealistic_v33.safetensors [82b0d085], dalcefo_v4.safetensors [425952fe], deliberate_v2.safetensors [10ec4b29], deliberate_v3.safetensors [afd9d2d4], dreamlike-anime-1.0.safetensors [4520e090], dreamlike-diffusion-1.0.safetensors [5c9fd6e0], dreamlike-photoreal-2.0.safetensors [fdcf65e7], dreamshaper_6BakedVae.safetensors [114c8abb], dreamshaper_7.safetensors [5cf5ae06], dreamshaper_8.safetensors [9d40847d], edgeOfRealism_eorV20.safetensors [3ed5de15], EimisAnimeDiffusion_V1.ckpt [4f828a15], elldreths-vivid-mix.safetensors [342d9d26], epicphotogasm_xPlusPlus.safetensors [1a8f6d35], epicrealism_naturalSinRC1VAE.safetensors [90a4c676], epicrealism_pureEvolutionV3.safetensors [42c8440c], ICantBelieveItsNotPhotography_seco.safetensors [4e7a3dfd], indigoFurryMix_v75Hybrid.safetensors [91208cbb], juggernaut_aftermath.safetensors [5e20c455], lofi_v4.safetensors [ccc204d6], lyriel_v16.safetensors [68fceea2], majicmixRealistic_v4.safetensors [29d0de58], mechamix_v10.safetensors [ee685731], meinamix_meinaV9.safetensors [2ec66ab0], meinamix_meinaV11.safetensors [b56ce717], neverendingDream_v122.safetensors [f964ceeb], openjourney_V4.ckpt [ca2f377f], pastelMixStylizedAnime_pruned_fp16.safetensors [793a26e8], portraitplus_V1.0.safetensors [1400e684], protogenx34.safetensors [5896f8d5], Realistic_Vision_V1.4-pruned-fp16.safetensors [8d21810b], Realistic_Vision_V2.0.safetensors [79587710], Realistic_Vision_V4.0.safetensors [29a7afaa], Realistic_Vision_V5.0.safetensors [614d1063], Realistic_Vision_V5.1.safetensors [a0f13c83], redshift_diffusion-V10.safetensors [1400e684], revAnimated_v122.safetensors [3f4fefd9], rundiffusionFX25D_v10.safetensors [cd12b0ee], rundiffusionFX_v10.safetensors [cd4e694d], sdv1_4.ckpt [7460a6fa], v1-5-pruned-emaonly.safetensors [d7049739], v1-5-inpainting.safetensors [21c7ab71], shoninsBeautiful_v10.safetensors [25d8c546], theallys-mix-ii-churned.safetensors [5d9225a4], timeless-1.0.ckpt [7c4971d4], toonyou_beta6.safetensors [980f6b15]

Removed Providers

g4f/Provider/Llama.py
g4f/Provider/selenium/AItianhuSpace.py

Provider Updates and Improvements

g4f/Provider/Nexra.py (AttributeError: 'BingCreateImages' #2181 (comment))
- Rename API endpoint variables for clarity
- Simplify model handling by separating text and image models
- Update image generation to return URL instead of base64 data
- Improve error handling in JSON parsing
- Remove unused imports and streamline code structurey.

g4f/Provider/ReplicateHome.py
- Simplify model lists and versions into single dictionaries
- Add support for system messages and message history
- Implement streaming support for text models
- Update API endpoint and request handling
- Improve error handling and response parsing
- Add new models: 'yorickvp/llava-13b' and 'black-forest-labs/flux-schnell'
- Remove unused imports and simplify type hints

g4f/Provider/AiChatOnline.py
- Remove supports_gpt_35_turbo attribute
- Remove supports_message_history attribute

g4f/Provider/bing/conversation.py
- Change bundleVersion from 1.1690.0 to 1.1809.0 in create_conversation function for both Copilot and regular Bing URLs

g4f/Provider/Koala.py
- Change url to include '/chat' path
- Add separate api_endpoint for API requests
- Update default_model from gpt-3.5-turbo to gpt-4o-mini
- Adjust Referer and Origin headers in create_async_generator

g4f/Provider/Snova.py
- Remove cookinai/DonutLM-v1 from models list
- Remove donutlm-v' from model_aliases
- Add comment indicating error with ignos/Mistral-T5-7B-v1 model

g4f/Provider/PerplexityLabs.py
- Add model_aliases dictionary to map shortened model names
- Change default_model from llama-3.1-8b-instruct to llama-3.1-70b-instruct

g4f/Provider/Bixin123.py (Provider updates and optimizations across multiple modules #2220 (comment))
- Add method to generate random fingerprint for API requests
- Create generate_fingerprint() class method
- Update create_async_generator() to use dynamic fingerprint

g4f/Provider/HuggingFace.py, g4f/Provider/HuggingChat.py (Error "KeyError: 'conversationId'" when using CohereForAI/c4ai-command-r-plus model in HuggingChat #2221 | Provider updates and optimizations across multiple modules #2220 (comment))
- Remove meta-llama/Meta-Llama-3.1-405B-Instruct-FP8 from models list
- Update CohereForAI/c4ai-command-r-plus to c4ai-command-r-plus-08-2024
- Update corresponding model alias for command-r-plus
- Remove 01-ai/Yi-1.5-34B-Chat from models list
- Remove yi-1.5-34b from model_aliases

g4f/Provider/Blackbox.py (Provider updates and optimizations across multiple modules #2220 (comment))
- Simplify model configuration with a unified model_config dictionary
- Update API endpoint URL and request headers for better compatibility
- Enhance image generation support for ImageGenerationLV45LJp model
- Remove unused methods and simplify code structure
- Adjust request payload structure for consistency across models
- Improve error handling for image generation responses

Unified the functionality of g4f/Provider/Rocks.py and g4f/Provider/FluxAirforce.py classes into the g4f/Provider/Airforce.py class for both text and image generation.

Consolidated g4f/Provider/Rocks.py and g4f/Provider/FluxAirforce.py APIs under a new g4f/Provider/Airforce.py provider class.
Implemented create_async_generator to handle both text and image generation.
Added the generate_text and generate_image methods for specific API endpoints:
- generate_text for the /chat/completions endpoint.
- generate_image for the /v1/imagine2 endpoint.
Introduced the format_prompt utility function for text generation.
Maintained backward compatibility with proxy and other configuration.
Expanded model lists for various types of generation.

…and improve readability

…ate_conversation URLs

…| Updated g4f/Provider/__init__.py

…fault model

github-actions

How can I assist you today?

github-actions · 2024-09-11T14:28:12Z

g4f/Provider/AiChatOnline.py

@@ -12,10 +12,8 @@ class AiChatOnline(AsyncGeneratorProvider, ProviderModelMixin):
    url = "https://aichatonlineorg.erweima.ai"
    api_endpoint = "/aichatonline/api/chat/gpt"
    working = True


Consider removing unnecessary attributes.

github-actions · 2024-09-11T14:28:12Z

g4f/Provider/AiChatOnline.py

@@ -12,10 +12,8 @@
    url = "https://aichatonlineorg.erweima.ai"
    api_endpoint = "/aichatonline/api/chat/gpt"
    working = True
-    supports_gpt_35_turbo = True
    supports_gpt_4 = True
    default_model = 'gpt-4o-mini'


Suggest removing unnecessary attribute.

github-actions · 2024-09-11T14:28:12Z

g4f/Provider/Koala.py

@@ -10,7 +10,8 @@
 from ..requests import raise_for_status

 class Koala(AsyncGeneratorProvider, ProviderModelMixin):


Consider updating the variable name 'url' to be more specific now that it includes '/chat'.

github-actions · 2024-09-11T14:28:12Z

g4f/Provider/Koala.py

@@ -10,7 +10,8 @@
 from ..requests import raise_for_status

 class Koala(AsyncGeneratorProvider, ProviderModelMixin):
-    url = "https://koala.sh"
+    url = "https://koala.sh/chat"
+    api_endpoint = "https://koala.sh/api/gpt/"


The addition of 'api_endpoint' is good but ensure it aligns with the naming convention used in this module.

github-actions · 2024-09-11T14:28:12Z

g4f/Provider/Koala.py

@@ -10,7 +10,8 @@
 from ..requests import raise_for_status

 class Koala(AsyncGeneratorProvider, ProviderModelMixin):
-    url = "https://koala.sh"
+    url = "https://koala.sh/chat"
+    api_endpoint = "https://koala.sh/api/gpt/"
    working = True
    supports_message_history = True


Since 'supports_message_history' attribute is no longer needed, consider removing it to avoid confusion.

github-actions · 2024-09-11T14:28:15Z

g4f/Provider/bing/conversation.py

    else:
-        url = "https://www.bing.com/turing/conversation/create?bundleVersion=1.1690.0"
+        url = "https://www.bing.com/turing/conversation/create?bundleVersion=1.1809.0"


Similarly, the Bing URL has been updated, but please verify that this change does not affect existing functionality.

github-actions · 2024-09-11T14:28:15Z

g4f/Provider/bing/conversation.py

@@ -90,4 +90,4 @@
            response = await response.json()
            return response["result"]["value"] == "Success"
    except:


The exception handling is too broad. It would be better to specify the type of exceptions you expect to catch, which can help in debugging and maintaining the code.

github-actions · 2024-09-11T14:28:15Z

g4f/Provider/selenium/__init__.py

@@ -1,4 +1,3 @@
-from .AItianhuSpace import AItianhuSpace
 from .MyShell import MyShell


Consider adding a brief comment explaining the purpose of importing MyShell.

github-actions · 2024-09-11T14:28:15Z

g4f/Provider/selenium/__init__.py

@@ -2,3 +1,3 @@
 from .MyShell import MyShell
 from .PerplexityAi import PerplexityAi


No improvements needed.

github-actions · 2024-09-11T14:28:15Z

g4f/Provider/selenium/__init__.py

@@ -2,3 +1,3 @@
 from .MyShell import MyShell
 from .PerplexityAi import PerplexityAi
 from .Phind import Phind


No improvements needed.

TheFirstNoob · 2024-09-11T16:57:10Z

Hey! after the last update my guys and I found a few more problems that can technically be fixed
Bixin123 if calling qwen-turbo model request Billing paymant. I think need remove this model.

Support conversation DDG like: #2210
Need check 50 message limits and remove conversation and create new. x-vqd-4 variable

In GUI interface after use generating image we have error like: file not found (generated_image_123456ID) like.

Blackbox Images sometimes work like chat not like Image Gen... its site problem.
Need rework some codes for calling image model cause like

.generate.image
model=blackbox (or ImageGenerationLV45LJp)

Calling model error

Feature request:
This idea well be realised on GUI interface but we can use this for AI chatting
Add duckduckgo-search for add ability AI use internet. Its huge improvement for AI.
I try use him and work some strange but work. Like link answering correcting: https://pypi.org/project/duckduckgo-search/
Or google https://pypi.org/project/googlesearch-python/

TheFirstNoob · 2024-09-12T10:28:44Z

#2221

More infos like testing Hugging providers

Models and errors:

command-r-plus
And need update to latest: CohereForAI/c4ai-command-r-plus-08-2024 for HuggingChat

Using HuggingChat provider
CohereForAI/c4ai-command-r-plus
HuggingChat: KeyError: 'conversationId'
Using HuggingFace provider
HuggingFace: ResponseStatusError: Response 403: {"error":"The model CohereForAI/c4ai-command-r-plus is too large to be loaded automatically (207GB > 10GB). Please use Spaces (https://huggingface.co/spaces) or Inference Endpoints (https://huggingface.co/inference-endpoints)."}

LLaMa 3.1 405B

Using HuggingChat provider
meta-llama/Meta-Llama-3.1-405B-Instruct-FP8
HuggingChat: KeyError: 'conversationId'
Using HuggingFace provider
HuggingFace: RateLimitError: Response 429: Rate limit reached

Phi-3-mini (Microsoft)

Using HuggingChat provider
microsoft/Phi-3-mini-4k-instruct
2024-09-12 13:23:01 ERROR    src.log -> Error to send message: No provider found

Yi-1.5-34B

Using HuggingChat provider
01-ai/Yi-1.5-34B-Chat
HuggingChat: KeyError: 'conversationId'

mixtral-8x7b
Error conversation chatting response streaming:

91 + 1031 equals 1122.
User: 328
Assistant: 32 8 equals 256.
User: 81/9
Assistant: 81 divided by 9 equals 9.
User: 122
Assistant: 12 2 equals 24.
User: 14/2
Assistant: 14 divided by 2 equals 7.
User: 15/5
Assistant: 15 divided by 5 equals 3.
User: 22+11
Assistant: 22 + 11 equals 33.
User: 22-11
Assistant: 22 - 11 equals 11.
User: 2211
Assistant: 22 11 equals 242.
User: 22/11
Assistant: 22 divided by 11 equals 2.
User: 22/2
Assistant: 22 divided by 2 equals 11.

And more answers...

TheFirstNoob · 2024-09-12T10:34:04Z

More info like Web Accessing Feature request:
HuggingFace used GoogleSearch libs like. I create some base code for test and sent here some examples later

TheFirstNoob · 2024-09-12T10:58:53Z

Add provider for use dalle-2 like here: #2181 (comment)
Thats your code but in g4f libs doesnt have calling dalle-2 model. Think is Nexra provider.

TheFirstNoob · 2024-09-12T11:10:27Z

Request Image Provider: https://app.prodia.com/
Endpoints?: https://api.prodia.com/generate and url = f'https://api.prodia.com/job/{job_id}'
Thanks!

…pdate provider models and aliases

… aliases

TheFirstNoob · 2024-09-12T15:55:09Z

@kqlio67 Discard latest commits
HuggingChat simply fix:

response = session.get(f'https://huggingface.co/chat/conversation/{conversationId}/__data.json?x-sveltekit-invalidated=11',)
We have invalidated=01
Change to 11 and all fully work

CR+ model update need stay

Log working:

Using HuggingChat provider
CohereForAI/c4ai-command-r-plus-08-2024
{'conversationId': '66e30e782540b6395c328cd0'}

Upd:
Remove models:
405B model remove its pro sub request now
mixtral-8x7b model have conversation answering problem and leak other chats.
Phi model have problem to find provider. We get ID fine, Model are fine but idk why provider didnt work correctly
Log Phi:

Using RetryProvider provider and microsoft/Phi-3-mini-4k-instruct model
Using HuggingChat provider
microsoft/Phi-3-mini-4k-instruct
{'conversationId': '66e31196689f9c7f5810c35a'}
2024-09-12 19:06:50 ERROR    src.log -> Error to send message: No provider found

All other models a work correclty

TheFirstNoob · 2024-09-12T18:50:30Z

More user friendly testing:
I think for huggingchat need divide alias mixtral model for more configurate and user friendly

However, we have many alias models with a similar feature, but this can only be done for this provider, since it is made separately on the site for ease of separation.
This minor request... It's up to you to decide)

"mixtral-8x7b": "mistralai/Mixtral-8x7B-Instruct-v0.1",
"mixtral-8x7b": "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",

If you call Mixtral-8x7b you always receive Nous-Hermes-2-Mixtral-8x7B-DPO and cant receive mistralai/Mixtral-8x7B-Instruct-v0.1

TheFirstNoob · 2024-09-12T19:30:00Z

@kqlio67 You are best man! Thanks for you working! :)
Sorry if I'm just telling you what to do. I don't understand coding very well yet and I'm still learning it fully.
I guess I need to wait for the merge now and then watch and test it!
Once again, thank you very much for your work! <3

…rove image generation

kqlio67 · 2024-09-12T20:52:13Z

@kqlio67 You are best man! Thanks for you working! :)
Sorry if I'm just telling you what to do. I don't understand coding very well yet and I'm still learning it fully.
I guess I need to wait for the merge now and then watch and test it!
Once again, thank you very much for your work! <3

Don't worry, I'm a beginner in coding too and not very proficient yet. I'm also learning and improving my skills with each project. I believe the best way to learn coding is through practice. This way we better understand how everything works and learn from our mistakes.

Public contributions to various projects really help to improve coding skills. Thank you for your suggestions - they give me an opportunity to try something new, and I'm learning and developing as a coder through this as well.

TheFirstNoob · 2024-09-12T21:29:31Z

Prodia:

line 27 and 28:
Duplicate models id:

        'AOM3A3_orangemixs.safetensors [9600da17]',
        'AOM3A3_orangemixs.safetensors [9600da17]',

xtekky · 2024-09-13T07:33:14Z

Excellent, I'll check this later today and merge it

kqlio67 added 9 commits September 11, 2024 15:54

refactor(g4f/Provider/AiChatOnline.py): streamline class attributes

80e17bc

refactor(g4f/Provider/ReplicateHome.py): optimize model organization …

7d1fa2c

…and improve readability

chore(g4f/Provider/bing/conversation.py): update bundleVersion in cre…

fc9cb1f

…ate_conversation URLs

refactor(g4f/Provider/Koala.py): update URL structure and default model

b18156a

refactor(g4f/Provider/Snova.py): update model list and add error note

7d028e5

Removed g4f/Provider/selenium/AItianhuSpace.py g4f/Provider/Llama.py …

a4f7540

…| Updated g4f/Provider/__init__.py

Updated g4f/Provider/selenium/__init__.py

a53954e

feat(g4f/Provider/PerplexityLabs.py): add model aliases and update de…

890e789

…fault model

Updated g4f/models.py

cd5b916

github-actions bot reviewed Sep 11, 2024

View reviewed changes

kqlio67 mentioned this pull request Sep 11, 2024

AttributeError: 'BingCreateImages' #2181

Closed

feat(g4f/Provider/Bixin123.py): implement dynamic fingerprint generation

2e3d60a

kqlio67 added 2 commits September 12, 2024 17:55

refactor(g4f/Provider/HuggingFace.py, g4f/Provider/HuggingChat.py): u…

abefd7f

…pdate provider models and aliases

refactor(models): update HuggingChat, HuggingFace provider models and…

11f702e

… aliases

kqlio67 added 4 commits September 12, 2024 20:28

Added a new provider for generating images: g4f/Provider/Prodia.py

2aa514b

Bug fixes and improvements to HuggingChat and HuggingFace providers

abea4dd

Bug fixes and improvements to HuggingChat provider

75bd017

Fixing errors with models in the provider

ba5f6ef

kqlio67 added 3 commits September 12, 2024 21:52

Updated g4f/models.py

b12c619

new model added

e23ee1f

Updated g4f/models.py

abdb27a

refactor(g4f/Provider/Blackbox.py): streamline model handling and imp…

a15578e

…rove image generation

Updated g4f/Provider/Blackbox.py

786a75b

Fixing errors in the provider g4f/Provider/Prodia.py

29515b6

kqlio67 added 5 commits September 13, 2024 17:09

feat(providers): unify Rocks and FluxAirforce into Airforce provider

3e491c6

Updated g4f/models.py

5b58852

Fixing errors in g4f/Provider/Nexra.py

f13557f

refactor(api): update model configuration for Blackbox

5d28b04

refactor(ReplicateHome): update model handling and API interaction

d69372a

xtekky merged commit cc80f2d into xtekky:main Sep 15, 2024
1 check passed

XInTheDark mentioned this pull request Sep 16, 2024

Update changes from g4f XInTheDark/raycast-g4f#114

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Provider updates and optimizations across multiple modules #2220

Provider updates and optimizations across multiple modules #2220

kqlio67 commented Sep 11, 2024 •

edited

Loading

github-actions bot left a comment

github-actions bot Sep 11, 2024

github-actions bot Sep 11, 2024

github-actions bot Sep 11, 2024

github-actions bot Sep 11, 2024

github-actions bot Sep 11, 2024

github-actions bot Sep 11, 2024

github-actions bot Sep 11, 2024

github-actions bot Sep 11, 2024

github-actions bot Sep 11, 2024

github-actions bot Sep 11, 2024

TheFirstNoob commented Sep 11, 2024 •

edited

Loading

TheFirstNoob commented Sep 12, 2024 •

edited

Loading

TheFirstNoob commented Sep 12, 2024

TheFirstNoob commented Sep 12, 2024 •

edited

Loading

TheFirstNoob commented Sep 12, 2024

TheFirstNoob commented Sep 12, 2024 •

edited

Loading

TheFirstNoob commented Sep 12, 2024 •

edited

Loading

TheFirstNoob commented Sep 12, 2024

kqlio67 commented Sep 12, 2024

TheFirstNoob commented Sep 12, 2024

xtekky commented Sep 13, 2024

		@@ -10,7 +10,8 @@
		from ..requests import raise_for_status

		class Koala(AsyncGeneratorProvider, ProviderModelMixin):

		@@ -1,4 +1,3 @@
		from .AItianhuSpace import AItianhuSpace
		from .MyShell import MyShell

		@@ -2,3 +1,3 @@
		from .MyShell import MyShell
		from .PerplexityAi import PerplexityAi

Provider updates and optimizations across multiple modules #2220

Provider updates and optimizations across multiple modules #2220

Conversation

kqlio67 commented Sep 11, 2024 • edited Loading

New providers added

Removed Providers

Provider Updates and Improvements

github-actions bot left a comment

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

github-actions bot Sep 11, 2024

Choose a reason for hiding this comment

TheFirstNoob commented Sep 11, 2024 • edited Loading

TheFirstNoob commented Sep 12, 2024 • edited Loading

TheFirstNoob commented Sep 12, 2024

TheFirstNoob commented Sep 12, 2024 • edited Loading

TheFirstNoob commented Sep 12, 2024

TheFirstNoob commented Sep 12, 2024 • edited Loading

TheFirstNoob commented Sep 12, 2024 • edited Loading

TheFirstNoob commented Sep 12, 2024

kqlio67 commented Sep 12, 2024

TheFirstNoob commented Sep 12, 2024

xtekky commented Sep 13, 2024

kqlio67 commented Sep 11, 2024 •

edited

Loading

TheFirstNoob commented Sep 11, 2024 •

edited

Loading

TheFirstNoob commented Sep 12, 2024 •

edited

Loading

TheFirstNoob commented Sep 12, 2024 •

edited

Loading

TheFirstNoob commented Sep 12, 2024 •

edited

Loading

TheFirstNoob commented Sep 12, 2024 •

edited

Loading