LLama4 Attention: Update assertion msg (#12777)
This commit is contained in:
@@ -1022,7 +1022,7 @@ class ServerArgs:
|
||||
"triton",
|
||||
"trtllm_mha",
|
||||
"intel_xpu",
|
||||
}, "fa3, aiter, triton, trtllm_mha or intel_xpu is required for Llama4 model"
|
||||
}, f"fa3, aiter, triton, trtllm_mha or intel_xpu is required for Llama4 model but got {self.attention_backend}"
|
||||
if is_sm100_supported() and self.attention_backend is None:
|
||||
self.attention_backend = "trtllm_mha"
|
||||
logger.warning(
|
||||
|
||||
Reference in New Issue
Block a user