Post-training with Tinker

Official repository for the paper "LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code"

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

Fully open data curation for reasoning models

🤗 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 repository for FLAME-MoE: A Transparent End-to-End Research Platform for Mixture-of-Experts Language Models

A high-throughput and memory-efficient inference and serving engine for LLMs

Reward-guided Speculative Decoding (RSD) for efficiency and effectiveness.

A simplified implementation for experimenting with RLVR on GSM8K, This repository provides a starting point for exploring reasoning.

FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.

Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding

[MLSys 2024 Best Paper Award] AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration

Reduce end to end training time from days to hours (or hours to minutes), and energy requirements/costs by an order of magnitude using coresets and data selection.

🎨 Semantic segmentation models, datasets and losses implemented in PyTorch.

Deep neural network to extract intelligent information from invoice documents.

Gradient Starvation: A Learning Proclivity in Neural Networks

Project Done for the course on Data Analytics UE18CS312

Network Slimming (Pytorch) (ICCV 2017)