Stars
"LightRAG: Simple and Fast Retrieval-Augmented Generation"
Pip compatible CodeBLEU metric implementation available for linux/macos/win
Code for Paper (ReMax: A Simple, Efficient and Effective Reinforcement Learning Method for Aligning Large Language Models)
Code for Adam-mini: Use Fewer Learning Rates To Gain More https://arxiv.org/abs/2406.16793
Python code of Hilbert-Schmidt Independence Criterion
HuatuoGPT, Towards Taming Language Models To Be a Doctor. (An Open Medical GPT)
Gaussian mixture model for incomplete (missing or truncated) and noisy data
Well-documented Python demonstrations for spatial data analytics, geostatistical and machine learning to support my courses.
BackPACK - a backpropagation package built on top of PyTorch which efficiently computes quantities other than the gradient.
A Python library that helps data scientists to infer causation rather than observing correlation.
Root Cause Discovery: Root Cause Analysis of Failures in Microservices through Causal Discovery
This is the official repo for the paper "Zeroth-Order Optimization Meets Human Feedback: Provable Learning via Ranking Oracles", Tang et al. https://arxiv.org/abs/2303.03751
基于注意力机制的少量样本故障诊断 pytorch
Active Bayesian Causal Inference (Neurips'22)
Causal Effect Engine is a Golang package for causal inference
Light-weighted code for Orthogonal Additive Gaussian Processes
This is a code demo for post-nonlinear mixture learning under simplex structure.
A Python implementation of global optimization with gaussian processes.
Combustion engine simulator that generates realistic audio.