Katelyn Li
Katelyn Li
Hi @giovp, Thank you so much for your reply, I've tried the above suggestions and it helped! Unfortunately, the highest batch size I could set without running into gpu memory...
Thank you so much! I increased the `batch_size` to `8192` and changed the `threshold` to `0.01`. It is still taking quite some time, and I think it would go much...
```(base) [katelyn.li@gpusrv42 ~]$ nvidia-smi Thu May 25 16:50:05 2023 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 525.60.13 Driver Version: 525.60.13 CUDA Version: 12.0 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr....
Thank you! This makes sense, and I'll try this and report back.