[CUDA] Support 3/5/6-bit quants in QMV (#3236)

This commit is contained in:
Cheng
2026-03-10 19:09:48 +09:00
committed by GitHub
parent 8d022bcb86
commit 9d03a1b0d9
3 changed files with 108 additions and 7 deletions
-1
View File
@@ -43,7 +43,6 @@ cuda_skip = {
"TestQuantized.test_qmm_jvp",
"TestQuantized.test_qmm_shapes",
"TestQuantized.test_qmm_vjp",
"TestQuantized.test_qmv",
"TestQuantized.test_fp_qvm",
"TestQuantized.test_qvm",
"TestQuantized.test_qvm_splitk",