[Fusion] Support 3D activation quantization fusion and enable it for Pixtral (#59612)
Signed-off-by: Jakub Zakrzewski <jzakrzewski@nvidia.com> Co-authored-by: Codex <noreply@openai.com> Co-authored-by: Misha Goin <mgoin64@gmail.com>
This commit is contained in:
co-authored by
Codex
Misha Goin
parent
dd30786fef
commit
19c0bed584
@@ -21,6 +21,7 @@ from vllm.model_executor.kernels.linear.nvfp4.base import (
|
||||
NvFp4LinearLayerConfig,
|
||||
)
|
||||
from vllm.model_executor.kernels.linear.nvfp4.flashinfer import (
|
||||
FlashInferCuteDslNvFp4LinearKernel,
|
||||
FlashInferCutlassNvFp4LinearKernel,
|
||||
FlashInferTrtllmNvFp4LinearKernel,
|
||||
)
|
||||
@@ -59,6 +60,7 @@ from vllm.platforms import current_platform
|
||||
SUPPORTING = {
|
||||
CutlassFP8ScaledMMLinearKernel,
|
||||
FlashInferFP8ScaledMMLinearKernel,
|
||||
FlashInferCuteDslNvFp4LinearKernel,
|
||||
FlashInferCutlassNvFp4LinearKernel,
|
||||
PerTensorTorchFP8ScaledMMLinearKernel,
|
||||
AiterHipbMMPerTokenFp8ScaledMMLinearKernel,
|
||||
|
||||
Reference in New Issue
Block a user