Base model: Qwen/Qwen3.5-2B
Base revision: 15852e8c16360a2fea060d615a32b45270f8a8fc
Original model and license: Qwen Team

FineEnvs fine-tuned model.safetensors using TRL Async GRPO.
Checkpoint: 700; release revision: main.
Configuration, generation settings, tokenizer serialization and chat template
are the saved training-checkpoint versions and may differ from the base release.
The original license is retained in LICENSE.
Experiment: https://huggingface.co/spaces/AdithyaSK/multi-harness-rl