https://github.com/mlabonne/mlabonne

https://github.com/mlabonne/mlabonne

Last synced: 2 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/mlabonne/mlabonne
• Owner: mlabonne
• Created: 2024-01-14T00:53:45.000Z (almost 2 years ago)
• Default Branch: main
• Last Pushed: 2025-03-19T20:32:27.000Z (7 months ago)
• Last Synced: 2025-07-08T08:46:10.570Z (4 months ago)
• Size: 32.2 KB
• Stars: 22
• Watchers: 2
• Forks: 3
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          


  


    𝕏 Follow me on X • 

    🤗 Hugging Face • 

    💻 Blog • 

    📙 LLM Engineer's Handbook

  






Hi, I'm a Machine Learning Scientist, Author, Blogger, and LLM Developer.

## 💼 Projects

* [**The LLM Course**](https://github.com/mlabonne/llm-course): A popular curated list of resources to get into LLMs (>57k ⭐).

* [**LLM Engineer's Handbook**](https://github.com/PacktPublishing/LLM-Engineers-Handbook): My book about LLM engineering with fine-tuning, RAG, data, evaluation, deployment, etc.

* [**Hands-on GNNs**](https://github.com/PacktPublishing/Hands-On-Graph-Neural-Networks-Using-Python): My book about designing and implementing graph neural networks in practice.

* [**LLM Datasets**](https://github.com/mlabonne/llm-datasets): Curated list of high-quality datasets for LLM fine-tuning.

* [**LLM Tools**](https://github.com/mlabonne/llm-course?tab=readme-ov-file#tools): Automate LLM pipelines with Colab notebooks like [LLM AutoEval](https://github.com/mlabonne/llm-autoeval), [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing), [LazyAxolotl](https://colab.research.google.com/drive/1TsDKNo2riwVmU55gjuBgB1AXVtRRfRHW?usp=sharing), and [AutoQuant](https://colab.research.google.com/drive/1b6nqC7UZVt8bx4MksX7s656GXPM-eWw4?usp=sharing).

## 🤗 New Models

* [**Liquid Foundation Models**](https://www.liquid.ai/liquid-foundation-models): My work at Liquid AI is to post-train our own pre-trained LLMs with a custom architecture. [[Playground]](https://playground.liquid.ai/chat)

* [**Abliterated models**](https://huggingface.co/collections/mlabonne/abliteration-66bf9a0f9f88f7346cb9462f): Collection of abliterated models to remove refusals. [[Article]](https://huggingface.co/blog/mlabonne/abliteration)

* [**NeuralDaredevil-8B**](https://huggingface.co/mlabonne/NeuralDaredevil-8B-abliterated): Uncensored models with the highest MMLU score in the 8B category.

## 💀 Old models

* [**AlphaMonarch-7B**](https://huggingface.co/mlabonne/AlphaMonarch-7B): Top performer in terms of reasoning + conversational abilities on a variety of benchmarks. [[Demo]](https://huggingface.co/spaces/mlabonne/AlphaMonarch-7B-GGUF-Chat)

* [**NeuralBeagle14-7B**](https://huggingface.co/mlabonne/NeuralBeagle14-7B): The most powerful 7B model (rank 10 on the *entire* Open LLM Leaderboard). [[Demo]](https://huggingface.co/spaces/mlabonne/NeuralBeagle14-7B-GGUF-Chat)

* [**Phixtral**](https://huggingface.co/mlabonne/phixtral-2x2_8): Novel Mixture of Experts architecture with phi-2 models. [[Demo]](https://huggingface.co/spaces/mlabonne/phixtral-chat)

* [**Beyonder-4x7B-v3**](https://huggingface.co/mlabonne/Beyonder-4x7B-v3): Mixture of Experts with four excellent fine-tuned Mistral-7b models. [[Demo]](https://huggingface.co/spaces/mlabonne/Beyonder-4x7B-v3-GGUF-Chat)

* [**NeuralHermes**](https://huggingface.co/mlabonne/NeuralHermes-2.5-Mistral-7B): A DPO fine-tuned version of OpenHermes (extremely cost-efficient). [[Demo]](https://huggingface.co/spaces/mlabonne/NeuralHermes-2.5-Mistral-7B-GGUF-Chat)