add 2/3/5/6/7 CUDA support in AutoRound format (#52890)

Signed-off-by: Wenhua Cheng <wenhua.cheng@intel.com>
This commit is contained in:
Wenhua Cheng
2026-09-07 09:59:19 +08:00
committed by GitHub
parent 199cb9b964
commit 294fbb4f59
3 changed files with 152 additions and 2 deletions
@@ -34,9 +34,8 @@ class INCConfig(QuantizationConfig):
Repo: https://github.com/intel/neural-compressor
"""
SUPPORTED_BITS = {2, 3, 4, 5, 6, 7, 8}
DEFAULT_INT_PACKING_FORMAT = "auto_round:auto_gptq"
SUPPORTED_BITS = {2, 3, 4, 8}
SUPPORTED_DTYPES = {"int", "mx_fp", "fp"}
SUPPORTED_FORMATS = {
"auto_round:auto_gptq",