MoBA: Mixture of Block Attention for Long-Context LLMs

MLE-bench is a benchmark for measuring how well AI agents perform at machine learning engineering

Yelp Simulator for WWW'25 AgentSociety Challenge

Big five trait scores for 307,313 people from many different countries.

Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

Official inference framework for 1-bit LLMs

SWE-agent takes a GitHub issue and tries to automatically fix it, using your LM of choice. It can also be employed for offensive cybersecurity or competitive coding challenges. [NeurIPS 2024]

SWE-bench: Can Language Models Resolve Real-world Github Issues?

Official repo for the paper "Scaling Synthetic Data Creation with 1,000,000,000 Personas"

Repo for paper "Unleashing Cognitive Synergy in Large Language Models: A Task-Solving Agent through Multi-Persona Self-Collaboration"

[NeurIPS '23 Spotlight] Thought Cloning: Learning to Think while Acting by Imitating Human Thinking

Repo for "Smart Word Suggestions" (SWS) task and benchmark

Convert Machine Learning Code Between Frameworks

Pre-Training with Whole Word Masking for Chinese BERT（中文BERT-wwm系列模型）

Chinese version of GPT2 training code, using BERT tokenizer.

pycorrector is a toolkit for text error correction. 文本纠错，实现了Kenlm，T5，MacBERT，ChatGLM3，Qwen2.5等模型应用在纠错场景，开箱即用。

A C# based MediaPlayer desktop application.

Sequence to Sequence Learning with Keras

Four styles of encoder decoder model by Python, Theano, Keras and Seq2Seq

header only, dependency-free deep learning framework in C++14