llm

KO | EN

gitlite — search

#javascript #react #llm #go #golang #android #machine-learning #docker #java #windows #awesome #cli #pytorch #openai

(★ 15)

Fine-tuning, DPO, RLHF, RLAIF on LLMs - Qwen3, Zephyr 7B GPTQ with 4-Bit Quantization, Mistral-7B-GPTQ

llm 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# LLM Training and Fine-tuning Projects

This repository contains various LLM training implementations using different optimization approaches.

## Projects by Training Approach

### Self-Supervised Alignment

* **[MLX-GRPO: Group Relative Policy Optimization](https://github.com/bayjarvis/llm/tree/main/mlx-grpo)** - Complete GRPO implementation for Apple Silicon using MLX framework with Qwen3-0.6B model support. Uses group comparisons for alignment without requiring human feedback data.

### Human Feedback-Based Training

* **[Harnessing Zephyr's Breeze: DPO Training on Mistral-7B-GPTQ](https://github.com/bayjarvis/llm/tree/main/mistral/dpo)** - Direct Preference Optimization for language model alignment using human preference datasets on quantized models.

### Supervised Fine-tuning

* **[Fine-tuning Zephyr 7B GPTQ with 4-Bit Quantization](https://github.com/bayjarvis/llm/tree/main/zephyr/finetune_gptq)** - Custom data fine-tuning with 4-bit quantization for efficient inference and deployment.

### Architectural Implementations

* **[Mixture of Experts (MoE) in PyTorch](https.github.com/bayjarvis/llm/tree/main/moe)** - A from-scratch implementation of a sparse Mixture of Experts layer in PyTorch, demonstrating a key technique for building large, efficient language models.

llm

llm 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

llm

llm 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README