search

intel / neural-compressor

SOTA low-bit LLM quantization (INT8/FP8/INT4/FP4/NF4) & sparsity; leading model compression techniques on TensorFlow, PyTorch, and ONNX Runtime
https://intel.github.io/neural-compressor/
Apache License 2.0
2.18k stars 252 forks source link

Remove Gelu Fusion for TF Newapi #1886

Closed zehao-intel closed 3 months ago

zehao-intel commented 3 months ago

Type of Change

bug fix API not changed

Description

TFDO removed Gelu op after some version of gnr-base. We need to disable the fusion of Gelu in pre-optimize path for TF newapi. The related tests should also be removed.

How has this PR been tested?

PreCI

Dependency Change?

No

github-actions[bot] commented 3 months ago

⚡ Required checks status: All passing 🟢

Groups summary

🟢 Code Scan Tests workflow | Check ID | Status | Error details | | | -------- | ------ | ---- | --- | | [Code-Scan](https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_build/results?buildId=31776) | success | | ✅ | | [Code-Scan (Bandit Code Scan Bandit)](https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_build/results?buildId=31776&view=logs&jobId=343c57fa-283e-589b-e772-0a0553c93e53) | success | | ✅ | | [Code-Scan (DocStyle Code Scan DocStyle)](https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_build/results?buildId=31776&view=logs&jobId=c1e234ec-db76-5d40-e8f0-e1ad3ef905a3) | success | | ✅ | | [Code-Scan (Pylint Code Scan Pylint)](https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_build/results?buildId=31776&view=logs&jobId=454075da-6b11-57a5-edf2-4c5947924fa8) | success | | ✅ | These checks are required after the changes to `neural_compressor/tensorflow/quantization/utils/graph_rewriter/generic/fuse_gelu.py`.
🟢 Unit Tests 3x-TensorFlow workflow | Check ID | Status | Error details | | | -------- | ------ | ---- | --- | | [UT-3x-TensorFlow](https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_build/results?buildId=31778) | success | | ✅ | | [UT-3x-TensorFlow (Coverage Compare CollectDatafiles)](https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_build/results?buildId=31778&view=logs&jobId=71384379-497b-5787-1f51-cc2e0f831d78) | success | | ✅ | | [UT-3x-TensorFlow (Unit Test 3x TensorFlow Unit Test 3x TensorFlow)](https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_build/results?buildId=31778&view=logs&jobId=45daf969-8bc5-5011-3358-6610c803470a) | success | | ✅ | | [UT-3x-TensorFlow (Unit Test 3x TensorFlow baseline Unit Test 3x TensorFlow baseline)](https://dev.azure.com/lpot-inc/b7121868-d73a-4794-90c1-23135f974d09/_build/results?buildId=31778&view=logs&jobId=acbfe8aa-f0a9-5aa1-b544-bdcc7f0e36c4) | success | | ✅ | These checks are required after the changes to `neural_compressor/tensorflow/quantization/utils/graph_rewriter/generic/fuse_gelu.py`, `test/3x/tensorflow/quantization/ptq/newapi/test_graph_fuse_gelu_newapi.py`.

Thank you for your contribution! 💜

Note This comment is automatically generated and will be updates every 180 seconds within the next 6 hours. If you have any other questions, contact chensuyue or XuehaoSun for help.