iHore/vLLM
8
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

[CI][Bugfix] Correct the Laguna DFlash acceptance-length reference (#57647)

Browse Source 
Signed-off-by: Oxana Korzh <okorzh@amd.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: Andreas Karatzas <akaratza@amd.com>
This commit is contained in:
Oxana Korzh
2026-09-19 01:44:48 +00:00
committed by GitHub
co-authored by Claude Opus 5 Andreas Karatzas
parent 468663a5e5
commit 038cb14bd7
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/v1/e2e/spec_decode/acceptance_rates/dflash/test_dflash.py
+1 -1
View File
@@ -53,7 +53,7 @@ LAGUNA_DFLASH_NVFP4 = DFlashCorrectnessConfig(
model="poolside/Laguna-XS-2.1-NVFP4",
draft_model="poolside/Laguna-XS-2.1-DFlash-NVFP4",
expected_accuracy=0.7, # Standard GSM8K sanity floor.
expected_acceptance_len=3.55 * 0.9,
expected_acceptance_len=3.05 * 0.9,
num_speculative_tokens=15,
max_model_len=8192,
max_num_seqs=32,