use flashinfer.sampling (#18696)

This commit is contained in:
pansicheng
2026-02-26 10:02:38 +08:00
committed by GitHub
parent 2739d7df62
commit 2ad475b4ed
8 changed files with 14 additions and 431 deletions
-1
View File
@@ -329,7 +329,6 @@ set(SOURCES
"${repo-flashinfer_SOURCE_DIR}/csrc/norm.cu"
"${repo-flashinfer_SOURCE_DIR}/csrc/renorm.cu"
"${repo-flashinfer_SOURCE_DIR}/csrc/sampling.cu"
"${repo-flash-attention_SOURCE_DIR}/csrc/flash_attn/src/flash_fwd_sparse_hdim128_bf16_causal_sm80.cu"
"${repo-flash-attention_SOURCE_DIR}/csrc/flash_attn/src/flash_fwd_sparse_hdim128_bf16_sm80.cu"