mirror of
https://github.com/zenlm/gym.git
synced 2026-11-04 17:11:44 +00:00
- Create demo_training.py for full training workflow - Add simple_demo.py showing GSPO algorithm concepts - Demonstrate group normalization and sequence-level optimization - Show memory savings (40-60% vs PPO) - Ready for local training with Qwen3 models Run locally with: $ python simple_demo.py $ gym-cli train --stage gspo --config configs/gspo_qwen3_4b_nano.yaml