remove ipex and ccl for xpu and cpu#42852
Conversation
upstreamed to PyTorch or kernels-community; ccl for xpu is upstreamed to PyTorch built-in xccl backend; ccl for cpu will use gloo Signed-off-by: Yao, Matrix <matrix.yao@intel.com>
Signed-off-by: Matrix YAO <matrix.yao@intel.com>
|
@jiqing-feng , pls help review, thx. |
|
LGTM. |
|
pending on auto-round issue intel/auto-round#1154 to make the auto-round case pass. |
|
@SunMarc , pls help review, this is to fully remove ccl-backend and ipex from |
|
View the CircleCI Test Summary for this PR: https://huggingface.co/spaces/transformers-community/circle-ci-viz?pr=42852&sha=3eac2a |
SunMarc
left a comment
SunMarc
left a comment
There was a problem hiding this comment.
Thanks for the clean up ! Just a small nit
| output = quantized_model.generate(**input_ids, max_new_tokens=40, do_sample=False) | ||
| self.assertIn(self.tokenizer.decode(output[0], skip_special_tokens=True), self.EXPECTED_OUTPUTS) | ||
|
|
||
| @require_intel_extension_for_pytorch |
There was a problem hiding this comment.
maybe we can skip those tests or require a specific tag to be enabled ? It will be too slow on our CI to run the model on our cpu.
| inputs = tokenizer(text, return_tensors="pt").to(model.device) | ||
| tokenizer.decode(model.generate(**inputs, max_new_tokens=5)[0]) | ||
|
|
||
| @require_intel_extension_for_pytorch |
|
the CI failure is not brought by this PR |
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
|
[For maintainers] Suggested jobs to run (before merge) run-slow: autoround, gptq |
5 participants
remove ipex and ccl for xpu and cpu, since all ipex optimizations are upstreamed to PyTorch or kernels-community; ccl for xpu is upstreamed to PyTorch built-in xccl backend; ccl for cpu will use gloo