Vibrant Labs ORG

Supercharge Your LLM Application Evaluations 🚀

Ragas is a Python toolkit for automatically scoring the quality of AI-powered apps, measuring whether answers are accurate and grounded in source material, with built-in test data generation so you can start evaluating without a pre-made test set.

Research notes and extra resources for all the work at explodinggradients.com

Reward Model framework for LLM RLHF

Supervised instruction finetuning for LLM with HF trainer and Deepspeed

Code and Data for Evaluating the Evaluators

Cloning Bench

Unify Efficient Fine-tuning of 100+ LLMs

EnterpriseOps-Gym verified and improved

Growing RL environments from simulated worlds — an experiment in scalable environment design for LLM agents.

Experimental Ragas Evaluation Library

Repo to add code and data for synthetic test data generation

No description.

No description.

A simple text to sql agent

No description.

No description.

No description.