schinger

pong_actor-critic

Trains an agent with (stochastic) Policy Gradients(actor-critic) on Pong. Uses OpenAI Gym.

book

FullLLM

Full stack LLM (Pre-training/finetuning, PPO(RLHF), Inference, Quant, etc.)

AlphaZero

Simplest AlphaZero Implementation

DiffusionModel

Implement Diffusion Model only by Pytorch and MLP

hello-world

llama2.c

Inference Llama 2 in one file of pure C

models

Models and examples built with TensorFlow

mosesdecoder

Moses, the machine translation system

Paddle

PArallel Distributed Deep LEarning

PPO-simplest

PPO in one file

pytorch-pretrained-BERT

The Big-&-Extending-Repository-of-Transformers: PyTorch pretrained models for Google's BERT, OpenAI GPT & GPT-2, Google/CMU Transformer-XL.