xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-05-26 04:57:52 +08:00
Code Issues Packages Projects Releases Wiki Activity

fixup

Browse Source 
Signed-off-by: Tyler Michael Smith <tysmith@redhat.com>
This commit is contained in:
Tyler Michael Smith 2025-06-20 15:56:05 +00:00
parent 39d5d33f8f
commit 21ffc7353a
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/model_executor/layers/fused_moe/batched_deep_gemm_moe.py
View File

@ -18,9 +18,9 @@ has_deep_gemm = importlib.util.find_spec("deep_gemm") is not None
def _silu_mul_fp8_quant_deep_gemm( def _silu_mul_fp8_quant_deep_gemm(
# Pointers ------------------------------------------------------------ # Pointers ------------------------------------------------------------
input_ptr, # 16-bit activations (E, T, 2*H) input_ptr, # 16-bit activations (E, T, 2*H)
y_q_ptr, # fp88 quantized activations (E, T, H) y_q_ptr, # fp8 quantized activations (E, T, H)
y_s_ptr, # 16-bit scales (E, T, G) y_s_ptr, # 16-bit scales (E, T, G)
counts_ptr, # int32 num tokens per expert (E) counts_ptr, # int32 num tokens per expert (E)
# Sizes --------------------------------------------------------------- # Sizes ---------------------------------------------------------------
H: tl.constexpr, # hidden dimension (per output) H: tl.constexpr, # hidden dimension (per output)