SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

no_cuda does not take effect in non distributed environment (#23795)

Browse Source 
Signed-off-by: Wang, Yi <yi.a.wang@intel.com>
This commit is contained in:
Wang, Yi
2023-05-26 22:47:51 +08:00
committed by GitHub
parent d61d747627
commit b7b729b38d
1 changed files with 3 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/transformers/training_args.py
View File

@@ -1684,7 +1684,9 @@ class TrainingArguments:
) )
device = torch.device("mps") device = torch.device("mps")
self._n_gpu = 1 self._n_gpu = 1
elif self.no_cuda:
device = torch.device("cpu")
self._n_gpu = 0
else: else:
# if n_gpu is > 1 we'll use nn.DataParallel. # if n_gpu is > 1 we'll use nn.DataParallel.
# If you only want to use a specific subset of GPUs use `CUDA_VISIBLE_DEVICES=0` # If you only want to use a specific subset of GPUs use `CUDA_VISIBLE_DEVICES=0`