This repository contains the official implementation of "FastVLM: Efficient Vision Encoding for Vision Language Models" - CVPR 2025

Example models using DeepSpeed

s1: Simple test-time scaling

Firefly: 大模型训练工具，支持训练Qwen2.5、Qwen2、Yi1.5、Phi-3、Llama3、Gemma、MiniCPM、Yi、Deepseek、Orion、Xverse、Mixtral-8x7B、Zephyr、Mistral、Baichuan2、Llma2、Llama、Qwen、Baichuan、ChatGLM2、InternLM、Ziya2、Vicuna、Bloom等大模型

The official repo of Qwen-VL (通义千问-VL) chat & pretrained large vision language model proposed by Alibaba Cloud.

OpenMMLab's next-generation platform for general 3D object detection.

NanoDet-Plus⚡Super fast and lightweight anchor-free object detection model. 🔥Only 980 KB(int8) / 1.8MB (fp16) and run 97FPS on cellphone🔥

Modeling, training, eval, and inference code for OLMo

[CVPR 2024] Real-Time Open-Vocabulary Object Detection

Simple Online Realtime Tracking with a Deep Association Metric

[ECCV 2022] ByteTrack: Multi-Object Tracking by Associating Every Detection Box

Solve Visual Understanding with Reinforced VLMs

Data processing for and with foundation models! 🍎 🍋 🌽 ➡️ ➡️🍸 🍹 🍷

Awesome Pretrained Chinese NLP Models，高质量中文预训练模型&大模型&多模态模型&大语言模型集合

OpenPCDet Toolbox for LiDAR-based 3D Object Detection.

A PyTorch Implementation of Single Shot MultiBox Detector

🚀 「大模型」1小时从0训练26M参数的视觉多模态VLM！🌏 Train a 26M-parameter VLM from scratch in just 1 hours!

Code for robust monocular depth estimation described in "Ranftl et. al., Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer, TPAMI 2022"

Vision agent

Count the MACs / FLOPs of your PyTorch model.

OpenMMLab Text Detection, Recognition and Understanding Toolbox

📚A curated list of Awesome LLM/VLM Inference Papers with Codes: Flash-Attention, Paged-Attention, WINT8/4, Parallelism, etc.🎉

Torchreid: Deep learning person re-identification in PyTorch.

⛹️ Pytorch ReID: A tiny, friendly, strong pytorch implement of person re-id / vehicle re-id baseline. Tutorial 👉https://github.com/layumi/Person_reID_baseline_pytorch/tree/master/tutorial

[CVPR 2023 Best Paper Award] Planning-oriented Autonomous Driving

Simple, online, and realtime tracking of multiple objects in a video sequence.

MedicalGPT: Training Your Own Medical GPT Model with ChatGPT Training Pipeline. 训练医疗大模型，实现了包括增量预训练(PT)、有监督微调(SFT)、RLHF、DPO、ORPO、GRPO。

The simplest, fastest repository for training/finetuning small-sized VLMs.