search

ChanglongJiangGit / A2J-Transformer

[CVPR 2023] Code for paper 'A2J-Transformer: Anchor-to-Joint Transformer Network for 3D Interacting Hand Pose Estimation from a Single RGB Image'
Apache License 2.0
85 stars 7 forks source link

About code test. #5

Open JT-Sun opened 1 year ago

JT-Sun commented 1 year ago

Hello, I tried the installation instructions to test your code, but no matter how many Gpus I set up it will only run on Gpus 0. Also when I use one 3090 as gpu, it said cuda out of memory. But your paper said you test on a single 2080ti gpu. Looking forward to your reply!

ChanglongJiangGit commented 1 year ago

Thanks for your attention. For the first question, maybe you can change the when running the script python test.py --gpu <your_gpu_ids>. For example, if you want to use gpu 0,1,2,3, you can change it to 0,1,2,3 or 0-3. For the second question, this kind of problem is usually caused by the you set in config.py is too large. When I tested on 2080ti, the batch-size I used is 12. Maybe this can help you!

JT-Sun commented 1 year ago

Thanks for your reply! Thanks for your great work!

ChanglongJiangGit commented 1 year ago

No ploblem~