SGLang is a fast serving framework for large language models and vision language models.

Python library implementing a trie data structure.

Python library implementing a trie data structure.

A comprehensive library for implementing LLMs, including a unified training pipeline and comprehensive model evaluation.

Summarize existing representative LLMs text datasets.

High-quality datasets, tools, and concepts for LLM fine-tuning.

AdaLoRA: Adaptive Budget Allocation for Parameter-Efficient Fine-Tuning (ICLR 2023).

LLMs interview notes and answers:该仓库主要记录大模型（LLMs）算法工程师相关的面试题和参考答案

A quick guide (especially) for trending instruction finetuning datasets

Awesome Knowledge-Distillation. 分类整理的知识蒸馏paper(2014-2021)。

CA-LoRA: Adapting Existing LoRA for Compressed LLMs to Enable Efficient Multi-Tasking on Personal Devices (COLM 2024)

Exploring Efficient Fine-Grained Perception of Multimodal Large Language Models

CoreNet: A library for training deep neural networks

📰 Must-read papers and blogs on Speculative Decoding ⚡️

[TMLR 2024] Efficient Large Language Models: A Survey

Visualizer for neural network, deep learning and machine learning models

Reaching LLaMA2 Performance with 0.1M Dollars

LLM training in simple, raw C/CUDA

Course to get into Large Language Models (LLMs) with roadmaps and Colab notebooks.

Code for "AnyGPT: Unified Multimodal LLM with Discrete Sequence Modeling"

Code for the ICLR 2023 paper "GPTQ: Accurate Post-training Quantization of Generative Pretrained Transformers".

🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.

An 8bit automated quantization conversion tool for the pytorch (Post-training quantization based on KL divergence)

Multi-Candidate Speculative Decoding

Cool Papers - Immersive Paper Discovery

中文版 llm-numbers

how to optimize some algorithm in cuda.