axolotl-ai-cloud / grpo_codeLinks
A fast, local, and secure approach for training LLMs for coding tasks using GRPO with WebAssembly and interpreter feedback.
☆32Updated 3 months ago
Alternatives and similar repositories for grpo_code
Users that are interested in grpo_code are comparing it to the libraries listed below
Sorting:
- Simple GRPO scripts and configurations.☆59Updated 5 months ago
- ☆52Updated 8 months ago
- Verifiers for LLM Reinforcement Learning☆65Updated 3 months ago
- QAlign is a new test-time alignment approach that improves language model performance by using Markov chain Monte Carlo methods.☆23Updated 3 months ago
- ☆40Updated 7 months ago
- Optimizing Causal LMs through GRPO with weighted reward functions and automated hyperparameter tuning using Optuna☆55Updated 5 months ago
- LLM reads a paper and produce a working prototype☆58Updated 3 months ago
- Simple repository for training small reasoning models☆33Updated 5 months ago
- Very minimal (and stateless) agent framework☆44Updated 6 months ago
- Testing paligemma2 finetuning on reasoning dataset☆18Updated 6 months ago
- Training an LLM to use a calculator with multi-turn reinforcement learning, achieving a **62% absolute increase in evaluation accuracy**.☆42Updated 2 months ago
- ☆48Updated 5 months ago
- Official homepage for "Self-Harmonized Chain of Thought" (NAACL 2025)☆91Updated 5 months ago
- Train your own SOTA deductive reasoning model☆99Updated 4 months ago
- ☆11Updated 11 months ago
- Data preparation code for CrystalCoder 7B LLM☆45Updated last year
- entropix style sampling + GUI☆26Updated 8 months ago
- ☆87Updated last year
- Repository containing the SPIN experiments on the DIBT 10k ranked prompts☆24Updated last year
- accompanying material for sleep-time compute paper☆97Updated 2 months ago
- ☆69Updated last month
- Score LLM pretraining data with classifiers☆55Updated last year
- [ACL 2025] Agentic Reward Modeling: Integrating Human Preferences with Verifiable Correctness Signals for Reliable Reward Systems☆96Updated last month
- Official repository for "BLEUBERI: BLEU is a surprisingly effective reward for instruction following"☆25Updated last month
- ☆57Updated last week
- A public implementation of the ReLoRA pretraining method, built on Lightning-AI's Pytorch Lightning suite.☆33Updated last year
- Submodular optimization for context engineering: query fan-out, text selection, passage reranking☆33Updated this week
- ☆49Updated last week
- Codebase accompanying the Summary of a Haystack paper.☆79Updated 9 months ago
- Anchored Preference Optimization and Contrastive Revisions: Addressing Underspecification in Alignment☆60Updated 10 months ago