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

[docs] make empty_cache device-agnostic (#34774)

Browse Source 
make device-agnostic
This commit is contained in:
Fanli Lin
2024-11-19 01:58:26 +08:00
committed by GitHub
parent 36759f3312
commit 8568bf1bcf
1 changed files with 4 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
docs/source/en/training.md
View File

@@ -287,9 +287,10 @@ model.fit(tf_dataset)
At this point, you may need to restart your notebook or execute the following code to free some memory: At this point, you may need to restart your notebook or execute the following code to free some memory:
```py ```py
from accelerate.utils.memory import clear_device_cache
del model del model
del trainer del trainer
torch.cuda.empty_cache() clear_device_cache()
``` ```
Next, manually postprocess `tokenized_dataset` to prepare it for training. Next, manually postprocess `tokenized_dataset` to prepare it for training.
@@ -364,8 +365,9 @@ Lastly, specify `device` to use a GPU if you have access to one. Otherwise, trai
```py ```py
>>> import torch >>> import torch
>>> from accelerate.test_utils.testing import get_backend
>>> device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu") >>> device, _, _ = get_backend() # automatically detects the underlying device type (CUDA, CPU, XPU, MPS, etc.)
>>> model.to(device) >>> model.to(device)
``` ```