buaabai

dify

Dify is an open-source LLM app development platform. Dify's intuitive interface combines AI workflow, RAG pipeline, agent capabilities, model management, observability features and more, letting you quickly go from prototype to production.

llama3

The official Meta Llama 3 GitHub site

yolov10

YOLOv10: Real-Time End-to-End Object Detection

Digital

A digital logic designer and circuit simulator.

matmulfreellm

Implementation for MatMul-free LM.

DoRA

[ICML2024 (Oral)] Official PyTorch implementation of DoRA: Weight-Decomposed Low-Rank Adaptation

ao

Custom data types and layouts for training and inference

HolisticTraceAnalysis

A library to analyze PyTorch traces.

BitBLAS

BitBLAS is a library to support mixed-precision matrix multiplications, especially for quantized LLM deployment.

fp6_llm

An efficient GPU support for LLM inference with x-bit quantization (e.g. FP6,FP5).

BiLLM

(ICML 2024) BiLLM: Pushing the Limit of Post-Training Quantization for LLMs

BitMat

An efficent implementation of the method proposed in "The Era of 1-bit LLMs"

LLaMA3-Quantization

A repository dedicated to evaluating the performance of quantizied LLaMA3 using various quantization methods..

Quest

[ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference

AutoSmoothQuant

An easy-to-use package for implementing SmoothQuant for LLMs

Awesome-LLM-Quantization

Awesome list for LLM quantization

SLAB

[ICML 2024] Official PyTorch implementation of "SLAB: Efficient Transformers with Simplified Linear Attention and Progressive Re-parameterized Batch Normalization"

ShiftAddLLM

ShiftAddLLM: Accelerating Pretrained LLMs via Post-Training Multiplication-Less Reparameterization

FPGA-QOI

FPGA-based QOI image compressor and decompressor in Verilog language. 基于FPGA的QOI图像压缩器和解压器。

APT

[ICML'24 Oral] APT: Adaptive Pruning and Tuning Pretrained Language Models for Efficient Training and Inference

Floating-Point-Adder

32 bit pipelined binary floating point adder using IEEE-754 Single Precision Format in Verilog

evol-q

Quantization in the Jagged Loss Landscape of Vision Transformers

QST

Quantized Side Tuning: Fast and Memory-Efficient Tuning of Quantized Large Language Models

BGEMM-CUDA

This is a repository of Binary General Matrix Multiply (BGEMM) by customized CUDA kernel. Thank FP6-LLM for the wheels!

retraining-free-quantization

RFQuant: Retraining-free Model Quantization via One-Shot Weight-Coupling Learning, CVPR (2024)

qattn

Efficient GPU kernels for mixed-precision Vision Transformers in Triton

Uint-Packing

dac_sdc_2023_champion

Ansor-AF-DS

This repository contains the figures, tables data and source code in the paper ICS'24: "Accelerated Auto-Tuning of GPU Kernels for Tensor Computations".

SelectiveFocus