A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 03:30 UTC
[FIX][Intel GPU]fix ipex flash_attn_varlen_func api missing parameter (#20348)
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
Kunshang Ji committed
Jul 2, 2025 at 13:26 UTC
27b8017636c57f9d0cd182fb4287b798c4a7ef28
1 file changed
+1
vllm/_ipex_ops.py
+1
@@ -270,6 +270,7 @@ class ipex_ops:
270
q_descale=None,
271
k_descale=None,
272
v_descale=None,
273
+ num_splits=0,
274
):
275
if cu_seqlens_k is None:
276
# cu_seqlens_k is not used in ipex kernel.