[Quant] Add SupportsQuant to phi3 and clip (#13104)

2025-02-15 22:28:33 -05:00
parent 80f63a3966
commit 12913d17ba
25 changed files with 67 additions and 13 deletions
--- a/vllm/model_executor/layers/quantization/gptq_marlin_24.py
+++ b/vllm/model_executor/layers/quantization/gptq_marlin_24.py
@@ -38,6 +38,7 @@ class GPTQMarlin24Config(QuantizationConfig):
        weight_bits: int,
        group_size: int,
    ) -> None:
+        super().__init__()
        quant_type = {
            4: scalar_types.uint4b8,
            8: scalar_types.uint8b128,