Unverified Commit 0f9318f7 authored by Yineng Zhang's avatar Yineng Zhang Committed by GitHub
Browse files

feat: update auto_choose_speculative_params (#9470)

parent 849957bc
......@@ -2411,8 +2411,12 @@ def auto_choose_speculative_params(self: ServerArgs):
if arch in ["LlamaForCausalLM"]:
# The default value for llama
return (5, 4, 8)
elif arch in ["DeepseekV3ForCausalLM", "DeepseekV2ForCausalLM"]:
# The default value for deepseek
elif arch in [
"DeepseekV3ForCausalLM",
"DeepseekV2ForCausalLM",
"GptOssForCausalLM",
]:
# The default value for deepseek and gpt-oss
return (3, 1, 4)
elif arch in ["Grok1ForCausalLM", "Grok1VForCausalLM"]:
return (5, 4, 8)
......
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