llmsearch

long_llama

LongLLaMA is a large language model capable of handling long contexts. It is based on OpenLLaMA a...

accelerated-pytorch-transformers-generation

digital_palace

My Digital Palace - A Personal Journal for Reflection - A place to store all my thoughts

chain-of-hindsight

Chain-of-Hindsight, A Scalable RLHF Method

Megatron-LM

Ongoing research training transformer models at scale

starcoder

Home of StarCoder: fine-tuning & inference!

self-rag

This includes the original implementation of SELF-RAG: Learning to Retrieve, Generate and Critiqu...

starcoder2

Home of StarCoder2!

multi_token

Embed arbitrary modalities (images, audio, documents, etc) into large language models.

ChatAbstractions

LangChain chat model abstractions for dynamic failover, load balancing, chaos engineering, and more!

TUPE

Transformer with Untied Positional Encoding (TUPE). Code of paper "Rethinking Positional Encoding...

sentiment-discovery

Unsupervised Language Modeling at scale for robust sentiment classification

plain-transformers

Transformer models implementation for training from scratch.

hybrid-discriminative-generative

Hybrid Discriminative-Generative Training via Contrastive Learning

textbook_quality

Generate textbook-quality synthetic LLM pretraining data