danvics/haiku.rag
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Raise FRAMES input budget: qa max_tokens 8192, judge 16384

Browse source
This commit is contained in:
Yiorgis Gozadinos 2026-07-26 18:56:00 +03:00
parent a57a73b6d0
commit 5e10846292
No known key found for this signature in database
1 changed files with 4 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
evaluations/configs/frames.yaml
View file

@ -28,7 +28,9 @@ qa:
provider: openai provider: openai
name: gemma4-26b name: gemma4-26b
base_url: http://vllm:11432/v1 base_url: http://vllm:11432/v1
max_tokens: 49152 # vLLM reserves max_tokens out of max_model_len; a large value starves
# the input budget and 400s long agentic contexts.
max_tokens: 8192
evaluations: evaluations:
judge: judge:
@ -36,4 +38,4 @@ evaluations:
name: RedHatAI/Qwen3.6-35B-A3B-NVFP4 name: RedHatAI/Qwen3.6-35B-A3B-NVFP4
base_url: http://vllm:11430/v1 base_url: http://vllm:11430/v1
temperature: 0.0 temperature: 0.0
max_tokens: 4096 max_tokens: 16384