Open
Conversation
…pe-ppo-or-grpo Refactor RL pipeline for GRPO grouping
- Add environment.yml, requirements.txt for reproducible env setup - Add start.sh launch script with VLLM_CUDART_SO_PATH for /proc-restricted containers - Add .gitignore for temp_data, logs, checkpoints, and data files - Fix optimization_config.py for 2-GPU small-scale training (CodeContests_200 + MBPP) - Fix openrlhf_deepspeed.py: fallback to torch.optim.AdamW when FusedAdam JIT fails - Fix trainer.py: handle empty training batches and packed metadata averaging Co-authored-by: Cursor <cursoragent@cursor.com>
…ging, and 8-GPU config - Add experiment directory auto-creation with config snapshots and result archiving - Separate TensorBoard logs into dedicated tb_logs/ dir, pass step index for proper tracking - Add per-step training metrics logging (policy_loss, kl_loss, clip_ratio, entropy, lr) - Extend reward.py with estimated reward statistics output - Update optimization_config for Qwen2.5-7B-Instruct on 8-GPU (4×TP2) setup - Add config variants (debug, paper-aligned) under optimization/configs/ - Improve .gitignore to exclude logs, experiment checkpoints, and temp data Made-with: Cursor
- analysis/: training curves, eval summaries, BoN robustness analysis - figures/: training visualization plots - scripts: generate_report.py, regenerate_figures.py - DESIGN_SELF_BOOTSTRAPPED_GRPO.md: self-supervised reward design Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.