Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

It takes too much time to create data #19

Open
arturia-Xayah opened this issue Sep 18, 2024 · 3 comments
Open

It takes too much time to create data #19

arturia-Xayah opened this issue Sep 18, 2024 · 3 comments

Comments

@arturia-Xayah
Copy link

Can you please tell me how long should it take to generate data?Or publish the processed data?

@pschaldenbrand
Copy link
Member

Hello,
Could you please provide a little more information on the issue? It would be great to see what command you are running, so I can see if it is a problem with parameters. I'd also like more details on what you mean by "generate" and "processed" data.

Paintings generally take about 5-30 minutes to plan using my setup with an NVIDIA 4090 depending on number of brush strokes and optimization iterations. If yours is taking much longer than this, I think it's likely that you are not running this on the GPU. The code will take exceedingly long to run on CPU.

Best,
Peter

@arturia-Xayah
Copy link
Author

Hello, Could you please provide a little more information on the issue? It would be great to see what command you are running, so I can see if it is a problem with parameters. I'd also like more details on what you mean by "generate" and "processed" data.

Paintings generally take about 5-30 minutes to plan using my setup with an NVIDIA 4090 depending on number of brush strokes and optimization iterations. If yours is taking much longer than this, I think it's likely that you are not running this on the GPU. The code will take exceedingly long to run on CPU.

Best, Peter

Thanks for your help. I use CUDA_VISIBLE_DEVICES=0 bash create_copaint_data.sh to generate data with NVIDIA 3090 and I haven't changed any of the code in it . The command in create_copaint_data.sh is:

python3 create_copaint_data.py \
        --use_cache \
        --cache_dir ../src/caches/sharpie_short_strokes \
        --materials_json ../materials_ink.json \
        --lr_multiplier 2.0  \
        --n_iters 700  \
        --max_strokes_added 400  \
        --min_strokes_added 10  \
        --ink \
        --render_height 168 \
        --output_parent_dir train_data/ink

I checked that the value of the variable device(in create_copaint_data.py) is cuda. The running process is shown in the following figure.

image

Besides,it occupied about 8GB of CUDA memory. My python version is 3.11 with CUDA 12.2 .

@pschaldenbrand
Copy link
Member

Thanks for that information, and sorry for the late reply (I was traveling).

The timings here seem accurate with my system. Each drawing takes quite a while to create. I left the script running for a day or two to generate the data for training CoFRIDA.

I could look into sharing the data I used to train it, but it is large and therefore difficult to host.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants