GPT-2 (124M) quality in 5B tokens

LLM101n: Let's build a Storyteller

llama3 implementation one matrix multiplication at a time

A framework for few-shot evaluation of language models.

Generate 3D objects conditioned on text or images

Use DuckDuckGo for free with GPT-3.5-Turbo.

newspaper3k is a news, full-text, and article metadata extraction in Python 3. Advanced docs:

Just the facts -- web page content extraction

Text-to-image samples collected for the evaluation of DALL-E 3 in the whitepaper.

OpenAI Baselines: high-quality implementations of reinforcement learning algorithms

PyTorch package for the discrete VAE used for DALL·E.

Code for the paper "Language Models are Unsupervised Multitask Learners"

Evals is a framework for evaluating LLMs and LLM systems, and an open-source registry of benchmarks.

Code and model for the paper "Improving Language Understanding by Generative Pre-Training"

Unify Efficient Fine-Tuning of 100+ LLMs

Scripts for fine-tuning Meta Llama3 with composable FSDP & PEFT methods to cover single/multi-node GPUs. Supports default & custom datasets for applications such as summarization and Q&A. Supportin…

The official Meta Llama 3 GitHub site

tiktoken is a fast BPE tokeniser for use with OpenAI's models.

A Native-PyTorch Library for LLM Fine-tuning

Complete Open Source and Modular solution for MMO

An Extensible Toolkit for Finetuning and Inference of Large Foundation Models. Large Models for All.

Benchmarking long-form factuality in large language models. Original code for our paper "Long-form factuality in large language models".

how to optimize some algorithm in cuda.