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

[tests] add XPU part to testing (#34778)

Browse Source 
add XPU part to testing

Signed-off-by: Lin, Fanli <fanli.lin@intel.com>
This commit is contained in:
Fanli Lin
2024-11-19 01:59:11 +08:00
committed by GitHub
parent 9568a9dfc5
commit e80a65ba4f
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/testing.md
View File

@@ -428,7 +428,7 @@ pytest --instafail
### To GPU or not to GPU ### To GPU or not to GPU
On a GPU-enabled setup, to test in CPU-only mode add `CUDA_VISIBLE_DEVICES=""`: On a GPU-enabled setup, to test in CPU-only mode add `CUDA_VISIBLE_DEVICES=""` for CUDA GPUs:
```bash ```bash
CUDA_VISIBLE_DEVICES="" pytest tests/utils/test_logging.py CUDA_VISIBLE_DEVICES="" pytest tests/utils/test_logging.py
@@ -441,10 +441,12 @@ second gpu if you have gpus `0` and `1`, you can run:
CUDA_VISIBLE_DEVICES="1" pytest tests/utils/test_logging.py CUDA_VISIBLE_DEVICES="1" pytest tests/utils/test_logging.py
``` ```
For Intel GPUs, use `ZE_AFFINITY_MASK` instead of `CUDA_VISIBLE_DEVICES` in the above example.
This is handy when you want to run different tasks on different GPUs. This is handy when you want to run different tasks on different GPUs.
Some tests must be run on CPU-only, others on either CPU or GPU or TPU, yet others on multiple-GPUs. The following skip Some tests must be run on CPU-only, others on either CPU or GPU or TPU, yet others on multiple-GPUs. The following skip
decorators are used to set the requirements of tests CPU/GPU/TPU-wise: decorators are used to set the requirements of tests CPU/GPU/XPU/TPU-wise:
- `require_torch` - this test will run only under torch - `require_torch` - this test will run only under torch
- `require_torch_gpu` - as `require_torch` plus requires at least 1 GPU - `require_torch_gpu` - as `require_torch` plus requires at least 1 GPU