Skip to content

GitLab

Unverified Commit ed242652 authored by Shiyan Deng's avatar Shiyan Deng Committed by GitHub
Browse files

[bug] Make sure get_modality_with_max_tokens is deterministic (#34533) 


Signed-off-by: default avatarShiyan Deng <dsy842974287@meta.com>
parent b37b6797
Hide whitespace changes
Inline Side-by-side
Showing with 1 addition and 1 deletion
vllm/multimodal/encoder_budget.py
View file @ ed242652
...@@ -181,7 +181,7 @@ class MultiModalBudget: ...@@ -181,7 +181,7 @@ class MultiModalBudget:
def get_modality_with_max_tokens(self) -> str: def get_modality_with_max_tokens(self) -> str:
mm_max_toks_per_item = self.mm_max_toks_per_item mm_max_toks_per_item = self.mm_max_toks_per_item
modality, _ = max(mm_max_toks_per_item.items(), key=lambda x: x[1]) modality, _ = max(mm_max_toks_per_item.items(), key=lambda x: (x[1], x[0]))
return modality return modality
......
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