karpathy/nanochat

This page introduces an external open-source repository. It is not an HDATF product.

What it does

A minimal codebase for training and chatting with your own LLM on a single GPU node, covering tokenization, pretraining, finetuning, evaluation and inference. One depth setting determines the other hyperparameters.

How it helps ATF

Only loosely related. It could be study material on how the language models behind Harness executors are trained and evaluated, not a component to reuse.

License

MIT Permissive. Commercial use and changes are allowed if the copyright notice is kept.

More in this category

• jundot/omlx
  A local LLM inference server for Macs with Apple Silicon, managed from the macOS menu bar. It uses continuous batching and keeps KV cache in memory and on SSD so past context can be reused across requests.
• maderix/ANE
  A research proof of concept that trains neural networks directly on Apple's Neural Engine through reverse-engineered private APIs, without CoreML training APIs, Metal or the GPU. The author states it is not a maintained framework.
• TheTom/turboquant_plus
  TurboQuant+ is a research implementation of TurboQuant that compresses the transformer KV cache using PolarQuant and Walsh-Hadamard rotation. It holds a Python reference implementation, validation papers and benchmark data.
• AlexsJones/llmfit
  llmfit checks a machine's CPU, RAM, GPU and VRAM and recommends which open-source LLMs and quantizations it can run. It offers a terminal UI, a web dashboard, a REST API, and local benchmarks of real tokens per second.
• google-ai-edge/gallery
  An app that showcases on-device generative AI and lets people try open-source models locally and offline on mobile devices. It includes chat with a thinking mode and agent skills that add tools such as Wikipedia.

Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.