Remove quantization for bias in pattern (#26415)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/26415
We do dynamic quantization for bias right now, remove this in pattern
Test Plan:
python test/test_jit.py 'TestJit.test_quant_fusion'
Imported from OSS
Differential Revision: D17465555
fbshipit-source-id: 5e229cbc6ae85ea4ce727b3479993d79747d7792
