Skip to content

GitLab

Unverified Commit 274d8e88 authored by Woosuk Kwon's avatar Woosuk Kwon Committed by GitHub
Browse files

[V1][Minor] Enhance SpecDecoding Metrics Log in V1 (#15902) 


Signed-off-by: default avatarWoosuk Kwon <woosuk.kwon@berkeley.edu>
parent 2039c630
Hide whitespace changes
Inline Side-by-side
Showing with 10 additions and 7 deletions
vllm/v1/spec_decode/metrics.py
View file @ 274d8e88
...@@ -47,13 +47,16 @@ class SpecDecodingMetrics: ...@@ -47,13 +47,16 @@ class SpecDecodingMetrics:
num_draft_tokens = np.sum(self.num_draft_tokens) num_draft_tokens = np.sum(self.num_draft_tokens)
num_accepted_tokens = np.sum(self.num_accepted_tokens) num_accepted_tokens = np.sum(self.num_accepted_tokens)
draft_acceptance_rate = (num_accepted_tokens / num_draft_tokens draft_acceptance_rate = (num_accepted_tokens / num_draft_tokens *
if num_draft_tokens > 0 else float("nan")) 100 if num_draft_tokens > 0 else float("nan"))
logger.info( logger.info(
"Speculative metrics: " "SpecDecoding metrics: "
"Draft acceptance rate: %.3f, " "Draft acceptance rate: %.1f%%, "
"Number of accepted tokens: %d, " "Accepted: %d tokens, "
"Number of draft tokens: %d, ", draft_acceptance_rate, "Drafted: %d tokens",
num_accepted_tokens, num_draft_tokens) draft_acceptance_rate,
num_accepted_tokens,
num_draft_tokens,
)
self.reset() self.reset()
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