nawnoes

ML-Papers-of-the-Week

🔥Highlighting the top ML papers every week.

gemma_pytorch

The official PyTorch implementation of Google's Gemma models

transformer-debugger

langroid

Harness LLMs with Multi-Agent Programming

schedule_free

Schedule-Free Optimization in PyTorch

penzai

A JAX research toolkit for building, editing, and visualizing neural networks.

simple-evals

awesome-llm-powered-agent

Awesome things about LLM-powered agents. Papers / Repos / Blogs / ...

nanotron

Minimalistic large language model 3D-parallelism training

SPIN

The official implementation of Self-Play Fine-Tuning (SPIN)

DataDreamer

DataDreamer: Prompt. Generate Synthetic Data. Train & Align Models.   🤖💤

LongBench

[ACL 2024] LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding

RLHF-Reward-Modeling

Recipes to train reward model for RLHF.

ringattention

Transformers with Arbitrarily Large Context

ring-flash-attention

Ring attention implementation with flash attention

textbook_quality

Generate textbook-quality synthetic LLM pretraining data

large_language_model_training_playbook

An open collection of implementation tips, tricks and resources for training large language models

InternEvo

BlackMamba

Code repository for Black Mamba

terashuf

terashuf shuffles multi-terabyte text files using limited memory

LongAlign

LongAlign: A Recipe for Long Context Alignment Encompassing Data, Training, and Evaluation

miracl

A large-scale multilingual dataset for Information Retrieval. Thorough human-annotations across 18 diverse languages.

LogicKor

한국어 언어모델 다분야 사고력 벤치마크

scalax

A simple library for scaling up JAX programs

RethinkTinyLM

[ICML'24] The official implementation of “Rethinking Optimization and Architecture for Tiny Language Models”

Inflection-Benchmarks

Public Inflection Benchmarks

ReST-MCTS

ReST-MCTS*: LLM Self-Training via Process Reward Guided Tree Search

prepacking

The source code of our work "Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models"

sparse_feedback

ko-rm-judge

Reward Model을 이용하여 언어모델의 답변을 평가하기