zwhong714 / PSFTLinks
PSFT is a trust-region–inspired fine-tuning objective that views SFT as a policy gradient method with constant advantages, constraining policy drift to stabilize training and improve generalization.
☆26Updated last month
Alternatives and similar repositories for PSFT
Users that are interested in PSFT are comparing it to the libraries listed below
Sorting:
- Official codebase for "GenPRM: Scaling Test-Time Compute of Process Reward Models via Generative Reasoning".☆81Updated 4 months ago
- Exploring the Limit of Outcome Reward for Learning Mathematical Reasoning☆190Updated 6 months ago
- xVerify: Efficient Answer Verifier for Reasoning Model Evaluations☆131Updated 5 months ago
- Pre-trained, Scalable, High-performance Reward Models via Policy Discriminative Learning.☆157Updated 2 weeks ago
- ☆138Updated last week
- End-to-End Reinforcement Learning for Multi-Turn Tool-Integrated Reasoning☆292Updated 2 weeks ago
- ☆211Updated 7 months ago
- CPPO: Accelerating the Training of Group Relative Policy Optimization-Based Reasoning Models (NeurIPS 2025)☆152Updated 3 weeks ago
- Implementation for the research paper "Enhancing LLM Reasoning via Critique Models with Test-Time and Training-Time Supervision".☆56Updated 10 months ago
- Towards a Unified View of Large Language Model Post-Training☆144Updated last month
- MiroRL is an MCP-first reinforcement learning framework for deep research agent.☆163Updated last month
- ☆297Updated 4 months ago
- The official repo of SynLogic: Synthesizing Verifiable Reasoning Data at Scale for Learning Logical Reasoning and Beyond☆165Updated 3 months ago
- OpenRFT: Adapting Reasoning Foundation Model for Domain-specific Tasks with Reinforcement Fine-Tuning☆149Updated 9 months ago
- L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning☆256Updated 4 months ago
- ☆356Updated 4 months ago
- Scaling Preference Data Curation via Human-AI Synergy☆113Updated 3 months ago
- ☆155Updated 4 months ago
- A light-weight tool for evaluating LLMs in rule-based ways.☆69Updated 3 months ago
- Official Implementation of ARPO: End-to-End Policy Optimization for GUI Agents with Experience Replay☆128Updated 4 months ago
- Rethinking RL Scaling for Vision Language Models: A Transparent, From-Scratch Framework and Comprehensive Evaluation Scheme☆143Updated 6 months ago
- ☆168Updated 5 months ago
- REverse-Engineered Reasoning for Open-Ended Generation☆69Updated last month
- Extrapolating RLVR to General Domains without Verifiers☆171Updated last month
- The Entropy Mechanism of Reinforcement Learning for Large Language Model Reasoning.☆340Updated 2 months ago
- MPO: Boosting LLM Agents with Meta Plan Optimization (EMNLP 2025 Findings)☆71Updated last month
- A research repo for experiments about Reinforcement Finetuning☆52Updated 6 months ago
- [EMNLP 2025] LightThinker: Thinking Step-by-Step Compression☆104Updated 5 months ago
- ☆78Updated last month
- ☆67Updated 3 months ago