Quantized Attention that achieves speedups of 2.1-3.1x and 2.7-5.1x compared to FlashAttention2 and xformers, respectively, without lossing end-to-end metrics across various models.

Code release for "Masked-attention Mask Transformer for Universal Image Segmentation"

3D-Graphics-Rendering-Cookbook, Second Edition

Development repository for the Triton language and compiler

CUDA on non-NVIDIA GPUs

IGEV++: Iterative Multi-range Geometry Encoding Volumes for Stereo Matching

Deploy your model with TensorRT quickly.

【编程随想】收藏的电子书清单（多个学科，含下载链接）

The Open Cookbook for Top-Tier Code Large Language Model

reCamera is an opensource camera platform

Occupancy grid mapping using Python - KITTI dataset

A GPU-accelerated and parallelized occupancy grid mapping algorithm based on pytorch.

Codes of MVSFormer++: Revealing the Devil in Transformer’s Details for Multi-View Stereo (ICLR2024)

A Two-stage Consensus Filtering for Real-time 3D Registration

[CVPR 2024 - Oral, Best Paper Award Candidate] Marigold: Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation

[CVPR 2023] Official PyTorch implementation of "Learning Rotation-Equivariant Features for Visual Correspondence"

Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities

Depth Any Video with Scalable Synthetic Data

A CPU+GPU Profiling library that provides access to timeline traces and hardware performance counters.

[IROS'24 Oral] A Fully Open-source and Compact Aerial Robot with Omnidirectional Visual Perception

🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.

Open Source H.264 Codec

Odometry application of the accurate distance field based on Gaussian Processes

Strengthened Pose Information for self-supervised monocular depth estimation. SPIdepth refines the pose network to improve depth prediction accuracy, achieving state-of-the-art results on benchmark…

Official Implementation of LOTUS: Diffusion-based Visual Foundation Model for High-quality Dense Prediction

ConDense backbone, weights, and evaluation code.

[CVPR'19] Single-Image Piece-wise Planar 3D Reconstruction via Associative Embedding

Code to easily try 30 (and growing) different image matching methods