thomasverelst/dynconv
Code for Dynamic Convolutions: Exploiting Spatial Sparsity for Faster Inference (CVPR2020)
This project helps computer vision engineers and researchers speed up image processing tasks. It takes standard image data and applies a specialized convolution technique, focusing only on crucial parts of the image. The result is faster inference for applications like object classification or human pose estimation.
128 stars. No commits in the last 6 months.
Use this if you are developing or deploying computer vision models and need to reduce computational cost and inference time, especially for image classification or human pose estimation tasks.
Not ideal if your primary goal is to train entirely new model architectures from scratch, rather than optimize existing ones for speed.
Stars
128
Forks
14
Language
Cuda
License
—
Category
Last pushed
Jan 17, 2022
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/thomasverelst/dynconv"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
shadow2496/VITON-HD
Official PyTorch implementation of "VITON-HD: High-Resolution Virtual Try-On via...
theFoxofSky/ddfnet
The official implementation of the CVPR2021 paper: Decoupled Dynamic Filter Networks
ChristophReich1996/Involution
PyTorch reimplementation of the paper "Involution: Inverting the Inherence of Convolution for...
maryam089/SDViT
Official repository for "Self-Distilled Vision Transformer for Domain Generalization" (ACCV-2022 ORAL)
ananthu-aniraj/pdiscoformer
[ECCV 2024 Oral] Official implementation of the paper "PDiscoFormer: Relaxing Part Discovery...