vllm.kernels.triton ¶
Triton kernel implementations.
Modules:
| Name | Description |
|---|---|
qkv_padded_fp8_quant | Stride-aware FP8 quantization with head_dim padding for ViT attention. |
vllm.kernels.triton ¶Triton kernel implementations.
Modules:
| Name | Description |
|---|---|
qkv_padded_fp8_quant | Stride-aware FP8 quantization with head_dim padding for ViT attention. |