[DSv4] Adding TRTLLM gen attention kernel (#43827)
This commit is contained in:
@@ -55,7 +55,8 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
|
||||
|
||||
// Horizontally-fused DeepseekV4-MLA: per-head RMSNorm + GPT-J RoPE for Q, and
|
||||
// GPT-J RoPE + UE8M0 FP8 quant + paged cache insert for KV, all in one
|
||||
// kernel launch. Registered in _C_stable_libtorch.
|
||||
// kernel launch. Registered in _C_stable_libtorch (incl. the FlashInfer V4
|
||||
// full-cache bf16/fp8 variants).
|
||||
|
||||
// Quantization ops
|
||||
#ifndef USE_ROCM
|
||||
|
||||
Reference in New Issue
Block a user