CoolFace
Modelpublic

AccelerateScience/Qwen3-14B-gwb-press-conference-grpo

sourceHugging Facegpl-3.0updated 3mo agoView on Hugging Face
0likes4downloads
Model Card

Run UID from grid search: 83e313ad e1

Validation Score: 0.302 [0.286, 0.313]

Generation Config:

  • —do_sample: True
  • —temperature: 0.7
  • —top_p: 1.0
  • —top_k = 0
  • —repetition_penalty: 1.0
  • —maxnewtokens: 1024