[Misc] Disambiguate quantized types via a new ScalarType (#6396)

2024-08-02 16:51:58 -04:00
parent b482b9a5b1
commit a8d604ca2a
29 changed files with 1111 additions and 356 deletions
--- a/tests/kernels/test_int8_quant.py
+++ b/tests/kernels/test_int8_quant.py
@@ -1,8 +1,6 @@
 import pytest
 import torch

-# ruff: noqa: F401
-import vllm._C
 from tests.kernels.quant_utils import ref_dynamic_per_token_quant
 from vllm._custom_ops import scaled_int8_quant