Skip to content

GitLab

Unverified Commit 84b30d9e authored by zifeitong's avatar zifeitong Committed by GitHub
Browse files

Set the default attention backend for GLM-4.5v to fa3 (#9245)

parent ff0cf51c
Hide whitespace changes
Inline Side-by-side
Showing with 1 addition and 0 deletions
python/sglang/srt/utils.py
View file @ 84b30d9e
...@@ -2345,6 +2345,7 @@ def is_fa3_default_architecture(hf_config): ...@@ -2345,6 +2345,7 @@ def is_fa3_default_architecture(hf_config):
"Qwen3ForCausalLM", "Qwen3ForCausalLM",
"Qwen3MoeForCausalLM", "Qwen3MoeForCausalLM",
"Glm4MoeForCausalLM", "Glm4MoeForCausalLM",
"Glm4vMoeForConditionalGeneration",
"Step3VLForConditionalGeneration", "Step3VLForConditionalGeneration",
} }
return architectures[0] in default_archs return architectures[0] in default_archs
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment