[AMD] Add triton awq_dequantize kernel to support AWQ on ROCm (#7661)

This commit is contained in:
Hubert Lu
2025-07-18 14:27:25 -07:00
committed by GitHub
parent c8f31042a8
commit 7750b91ca8
5 changed files with 530 additions and 3 deletions

View File

@@ -147,6 +147,7 @@ suites = {
# TestFile("test_vision_chunked_prefill.py", 175), # Disabled temporarily and track in #7701
TestFile("test_reasoning_parser.py", 5),
TestFile("test_rope_rocm.py", 3),
TestFile("test_awq_dequant.py", 2),
],
"per-commit-npu": [
TestFile("test_ascend_attention_backend.py", 400),