saymrwulf/transformers
1
0
Fork 0
mirror of https://github.com/saymrwulf/transformers.git synced 2026-05-14 20:58:08 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

[doc] fix some typos and add xpu to the testing documentation (#29894)

Browse source 
fix typo
This commit is contained in:
Fanli Lin 2024-03-28 17:42:49 +08:00 committed by GitHub
parent 22d159ddf9
commit 7c19fafe44
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 11 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

19
docs/source/en/testing.md
View file

@ -168,7 +168,7 @@ pytest -k "ada and not adam" tests/test_optimization.py
For example to run both `test_adafactor` and `test_adam_w` you can use:
```bash
pytest -k "test_adam_w or test_adam_w" tests/test_optimization.py
pytest -k "test_adafactor or test_adam_w" tests/test_optimization.py
```
Note that we use `or` here, since we want either of the keywords to match to include both.
@ -457,7 +457,7 @@ Let's depict the GPU requirements in the following table:
| n gpus | decorator |
|--------+--------------------------------|
|--------|--------------------------------|
| `>= 0` | `@require_torch` |
| `>= 1` | `@require_torch_gpu` |
| `>= 2` | `@require_torch_multi_gpu` |
@ -518,21 +518,21 @@ To run the test suite on a specific torch device add `TRANSFORMERS_TEST_DEVICE="
TRANSFORMERS_TEST_DEVICE="cpu" pytest tests/utils/test_logging.py
```
This variable is useful for testing custom or less common PyTorch backends such as `mps`. It can also be used to achieve the same effect as `CUDA_VISIBLE_DEVICES` by targeting specific GPUs or testing in CPU-only mode.
This variable is useful for testing custom or less common PyTorch backends such as `mps`, `xpu` or `npu`. It can also be used to achieve the same effect as `CUDA_VISIBLE_DEVICES` by targeting specific GPUs or testing in CPU-only mode.
Certain devices will require an additional import after importing `torch` for the first time. This can be specified using the environment variable `TRANSFORMERS_TEST_BACKEND`:
```bash
TRANSFORMERS_TEST_BACKEND="torch_npu" pytest tests/utils/test_logging.py
```
Alternative backends may also require the replacement of device-specific functions. For example `torch.cuda.manual_seed` may need to be replaced with a device-specific seed setter like `torch.npu.manual_seed` to correctly set a random seed on the device. To specify a new backend with backend-specific device functions when running the test suite, create a Python device specification file in the format:
Alternative backends may also require the replacement of device-specific functions. For example `torch.cuda.manual_seed` may need to be replaced with a device-specific seed setter like `torch.npu.manual_seed` or `torch.xpu.manual_seed` to correctly set a random seed on the device. To specify a new backend with backend-specific device functions when running the test suite, create a Python device specification file `spec.py` in the format:
```
```python
import torch
import torch_npu
import torch_npu # for xpu, replace it with `import intel_extension_for_pytorch`
# !! Further additional imports can be added here !!
# Specify the device name (eg. 'cuda', 'cpu', 'npu')
# Specify the device name (eg. 'cuda', 'cpu', 'npu', 'xpu', 'mps')
DEVICE_NAME = 'npu'
# Specify device-specific backends to dispatch to.
@ -541,11 +541,10 @@ MANUAL_SEED_FN = torch.npu.manual_seed
EMPTY_CACHE_FN = torch.npu.empty_cache
DEVICE_COUNT_FN = torch.npu.device_count
```
This format also allows for specification of any additional imports required. To use this file to replace equivalent methods in the test suite, set the environment variable `TRANSFORMERS_TEST_DEVICE_SPEC` to the path of the spec file.
This format also allows for specification of any additional imports required. To use this file to replace equivalent methods in the test suite, set the environment variable `TRANSFORMERS_TEST_DEVICE_SPEC` to the path of the spec file, e.g. `TRANSFORMERS_TEST_DEVICE_SPEC=spec.py`.
Currently, only `MANUAL_SEED_FN`, `EMPTY_CACHE_FN` and `DEVICE_COUNT_FN` are supported for device-specific dispatch.
### Distributed training
`pytest` can't deal with distributed training directly. If this is attempted - the sub-processes don't do the right
@ -579,7 +578,7 @@ pytest -s tests/utils/test_logging.py
To send test results to JUnit format output:
```bash
py.test tests --junitxml=result.xml
pytest tests --junitxml=result.xml
```
### Color control

4
src/transformers/testing_utils.py
View file

@ -792,13 +792,13 @@ def require_torch_xpu(test_case):
def require_torch_multi_xpu(test_case):
"""
Decorator marking a test that requires a multi-XPU setup with IPEX and atleast one XPU device. These tests are
Decorator marking a test that requires a multi-XPU setup with IPEX and at least one XPU device. These tests are
skipped on a machine without IPEX or multiple XPUs.
To run *only* the multi_xpu tests, assuming all test names contain multi_xpu: $ pytest -sv ./tests -k "multi_xpu"
"""
if not is_torch_xpu_available():
return unittest.skip("test requires IPEX and atleast one XPU device")(test_case)
return unittest.skip("test requires IPEX and at least one XPU device")(test_case)
return unittest.skipUnless(torch.xpu.device_count() > 1, "test requires multiple XPUs")(test_case)