Sourab Mangrulkar's starred repositories
RWKV-LM
RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference, saves VRAM, fast training, "infinite" ctx_len, and free sentence embedding.
natural-instructions
Expanding natural instructions
pandarallel
A simple and efficient tool to parallelize Pandas operations on all available CPUs
cc2dataset
Easily convert common crawl to a dataset of caption and document. Image/text Audio/text Video/text, ...
annotated_deep_learning_paper_implementations
🧑🏫 60+ Implementations/tutorials of deep learning papers with side-by-side notes 📝; including transformers (original, xl, switch, feedback, vit, ...), optimizers (adam, adabelief, sophia, ...), gans(cyclegan, stylegan2, ...), 🎮 reinforcement learning (ppo, dqn), capsnet, distillation, ... 🧠
deep-rl-class
This repo contains the syllabus of the Hugging Face Deep Reinforcement Learning Course.
P-tuning-v2
An optimized deep prompt tuning strategy comparable to fine-tuning across scales and tasks
ml-stable-diffusion
Stable Diffusion with Core ML on Apple Silicon
diffusion-models-class
Materials for the Hugging Face Diffusion Models Course
flash-attention
Fast and memory-efficient exact attention
torchscale
Foundation Architecture for (M)LLMs
PromptPapers
Must-read papers on prompt-based tuning for pre-trained language models.
parallelformers
Parallelformers: An Efficient Model Parallelization Toolkit for Deployment
text-generation-inference
Large Language Model Text Generation Inference
pytorch_geometric
Graph Neural Network Library for PyTorch