fix: avoid non-deterministic error

ryanxingql · Oct 3, 2024 · 9f5f8ee · 9f5f8ee
1 parent 6aefea2
commit 9f5f8ee
Show file tree

Hide file tree

Showing 4 changed files with 5 additions and 5 deletions.
diff --git a/basicsr b/basicsr
diff --git a/doc.md b/doc.md
@@ -132,7 +132,7 @@ powerqe
 conda activate pqe
 
 #CUDA_VISIBLE_DEVICES=0 scripts/train.sh 1 options/train/ESRGAN/RRDBNet_DIV2K_LMDB_G1.yml --auto_resume
-[CUBLAS_WORKSPACE_CONFIG=:4096:8] CUDA_VISIBLE_DEVICES=<gpus> [PORT=<master_port>] scripts/train.sh <num_gpus> <cfg_path> [--auto_resume] [--debug] [--force_yml <key>=<value>]
+CUDA_VISIBLE_DEVICES=<gpus> [PORT=<master_port>] scripts/train.sh <num_gpus> <cfg_path> [--auto_resume] [--debug] [--force_yml <key>=<value>]
 ```
 
 - `auto_resume`: Automatically resume from the latest existing checkpoint.
@@ -147,7 +147,7 @@ conda activate pqe
 conda activate pqe
 
 #CUDA_VISIBLE_DEVICES=0 scripts/test.sh 1 options/test/ESRGAN/RRDBNet_DIV2K_LMDB_G1_latest.yml --force_yml path:pretrain_network_g=experiments/train_ESRGAN_RRDBNet_DIV2K_LMDB_G1/models/net_g_600000.pth
-[CUBLAS_WORKSPACE_CONFIG=:4096:8] CUDA_VISIBLE_DEVICES=<gpus> [PORT=<master_port>] scripts/test.sh <num_gpus> <cfg_path> [--force_yml <key>=<value>]
+CUDA_VISIBLE_DEVICES=<gpus> [PORT=<master_port>] scripts/test.sh <num_gpus> <cfg_path> [--force_yml <key>=<value>]
 ```
 
 - Most models support only single-GPU testing, even when multi-GPU testing is requested.

diff --git a/powerqe/test.py b/powerqe/test.py
@@ -16,7 +16,7 @@ def test_pipeline(root_path):
     if opt["reproduce"]:
         torch.backends.cudnn.benchmark = False
         torch.backends.cudnn.deterministic = True
-        torch.use_deterministic_algorithms(True)
+        torch.use_deterministic_algorithms(True, warn_only=True)
     else:
         torch.backends.cudnn.benchmark = True
         # torch.backends.cudnn.deterministic = True

diff --git a/powerqe/train.py b/powerqe/train.py
@@ -135,7 +135,7 @@ def train_pipeline(root_path):
     if opt["reproduce"]:
         torch.backends.cudnn.benchmark = False
         torch.backends.cudnn.deterministic = True
-        torch.use_deterministic_algorithms(True)
+        torch.use_deterministic_algorithms(True, warn_only=True)
     else:
         torch.backends.cudnn.benchmark = True
         # torch.backends.cudnn.deterministic = True