LAVT: Language-Aware Vision Transformer for Referring Image Segmentation

Overview

LAVT: Language-Aware Vision Transformer for Referring Image Segmentation

Where we are ?

12.27 目前和原论文仍有1%左右得差距,但已经力压很多SOTA了

ckpt__448_epoch_25.pth mIoU Overall IoU [email protected]
Refcoco val 70.743 71.671 82.26
Refcoco testA 73.679 74.772 -
Refcoco testB 67.582 67.339 -

12.29 45epoch的结果又上升了大约1%

ckpt__448_epoch_45.pth mIoU Overall IoU
Refcoco val 71.949 72.246
Refcoco testA 74.533 75.467
Refcoco testB 67.849 68.123

the pretrain model will be released soon

对原论文的复现

论文链接: https://arxiv.org/abs/2112.02244

官方实现: https://github.com/yz93/LAVT-RIS

Architecture

Features

  • 将不同模态feature的fusion提前到Image Encoder阶段

  • 思路上对这两篇论文有很多借鉴

    • Vision-Language Transformer and Query Generation for Referring Segmentation

    • Locate then Segment: A Strong Pipeline for Referring Image Segmentation

  • 采用了比较新的主干网络 Swin-Transformer

Usage

详细参数设置可以见args.py

for training

CUDA_VISIBLE_DEVICES=0,1,2,3 python -m torch.distributed.launch --nproc_per_node 4 --master_port 12345 main.py --batch_size 2 --cfg_file configs/swin_base_patch4_window7_224.yaml --size 448

for evaluation

CUDA_VISIBLE_DEVICES=4,5,6,7 python -m torch.distributed.launch --nproc_per_node 4 --master_port 23458 main.py --size 448 --batch_size 1 --resume --eval --type val --eval_mode cat --pretrain ckpt_448_epoch_20.pth --cfg_file configs/swin_base_patch4_window7_224.yaml

*.pth 都放在./checkpoint

for resume from checkpoint

CUDA_VISIBLE_DEVICES=0,1,2,3 python -m torch.distributed.launch --nproc_per_node 4 --master_port 12346 main.py --batch_size 2 --cfg_file configs/swin_base_patch4_window7_224.yaml --size 448 --resume --pretrain ckpt_448_epoch_10.pth

for dataset preparation

please get details from ./data/readme.md

Need to be finished

由于我在复现的时候,官方的code还没有出来,所以一些细节上的设置可能和官方code不同

  • Swin Transformer 我选择的是 swin_base_patch4_window12_384_22k.pth,具体代码可以参考官方代码 https://github.com/microsoft/Swin-Transformer/blob/main/get_started.md 原论文中的图像resize的尺寸是480*480,可是我目前基于官方的代码若想调到这个尺寸,总是会报错,查了一下觉得可能用object detection 的swin transformer的code比较好

    12.27 这个问题目前也已经得到了较好的解决,目前训练用的是 swin_base_patch4_window7_224_22k.pth, 输入图片的尺寸调整到448*448

    解决方案可以参考:

    https://github.com/microsoft/Swin-Transformer/issues/155

  • 原论文中使用的lr_scheduler是polynomial learning rate decay, 没有给出具体的参数手动设置了一下

    12.21 目前来看感觉自己设置的不是很好

    12.27 调整了一下设置,初始学习率的设置真的很重要,特别是根据batch_size 去scale你的 inital learning rate

  • 原论文中的batch_size=32,基于自己的实验我猜想应该是用了8块GPU,每一块的batch_size=4, 由于我第一次写DDP code,训练时发现,程序总是会在RANK0上给其余RANK开辟类似共享显存的东西,导致我无法做到原论文相同的配置,需要改进

  • 仔细观察Refcoco的数据集,会发现一个target会对应好几个sentence,training时我设计的是随机选一个句子,evaluate时感觉应该要把所有句子用上会更好,关于这一点我想了两种evaluate的方法

    目前eval 只能支持 batch_size=1

    • 将所有句子concatenate成为一个句子,送入BERT,Input 形式上就是(Image,cat(sent_1,sent_2,sent_3)) => model => pred

    实验发现这种eval_mode 下的mean IOU 会好不少, overall_IOU 也会好一点

    • 对同一张图片处理多次处理,然后将结果进行平均,Input 形式上就是 ((Image,sent_1),(Image,sent_2),(Image,sent_3)) => model => average(pred_1,pred_2,pred_3)

Visualization

详细见inference.ipynb

input sentences

  1. right girl
  2. closest girl on right

results

Failure cases study

AnalysisFailure.ipynb 提供了一个研究model不work的途径,主要是筛选了IoU < 0.5的case,并在这些case中着重查看了一下IoU < 0.10.4 < IoU < 0.5 的例子

目前我只看了一些有限的failure cases,做了如下总结

  • 模型对于similar,dense object在language guide下定位不精确
  • 模型对于language的理解不分主次
  • refcoco本身标记的一些问题
Owner
zichengsaber
CVer
zichengsaber
On Generating Extended Summaries of Long Documents

ExtendedSumm This repository contains the implementation details and datasets used in On Generating Extended Summaries of Long Documents paper at the

Georgetown Information Retrieval Lab 76 Sep 05, 2022
Source code for "OmniPhotos: Casual 360° VR Photography"

OmniPhotos: Casual 360° VR Photography Project Page | Video | Paper | Demo | Data This repository contains the source code for creating and viewing Om

Christian Richardt 144 Dec 30, 2022
Single Image Deraining Using Bilateral Recurrent Network (TIP 2020)

Single Image Deraining Using Bilateral Recurrent Network Introduction Single image deraining has received considerable progress based on deep convolut

23 Aug 10, 2022
All the essential resources and template code needed to understand and practice data structures and algorithms in python with few small projects to demonstrate their practical application.

Data Structures and Algorithms Python INDEX 1. Resources - Books Data Structures - Reema Thareja competitiveCoding Big-O Cheat Sheet DAA Syllabus Inte

Shushrut Kumar 129 Dec 15, 2022
Codes for our paper The Stem Cell Hypothesis: Dilemma behind Multi-Task Learning with Transformer Encoders published to EMNLP 2021.

The Stem Cell Hypothesis Codes for our paper The Stem Cell Hypothesis: Dilemma behind Multi-Task Learning with Transformer Encoders published to EMNLP

Emory NLP 5 Jul 08, 2022
[PAMI 2020] Show, Match and Segment: Joint Weakly Supervised Learning of Semantic Matching and Object Co-segmentation

Show, Match and Segment: Joint Weakly Supervised Learning of Semantic Matching and Object Co-segmentation This repository contains the source code for

Yun-Chun Chen 60 Nov 25, 2022
Implementation of a Transformer that Ponders, using the scheme from the PonderNet paper

Ponder(ing) Transformer Implementation of a Transformer that learns to adapt the number of computational steps it takes depending on the difficulty of

Phil Wang 65 Oct 04, 2022
Repository features UNet inspired architecture used for segmenting lungs on chest X-Ray images

Lung Segmentation (2D) Repository features UNet inspired architecture used for segmenting lungs on chest X-Ray images. Demo See the application of the

163 Sep 21, 2022
Deep Learning to Create StepMania SM FIles

StepCOVNet Running Audio to SM File Generator Currently only produces .txt files. Use SMDataTools to convert .txt to .sm python stepmania_note_generat

Chimezie Iwuanyanwu 8 Jan 08, 2023
A TensorFlow implementation of FCN-8s

FCN-8s implementation in TensorFlow Contents Overview Examples and demo video Dependencies How to use it Download pre-trained VGG-16 Overview This is

Pierluigi Ferrari 50 Aug 08, 2022
Implementation of the paper "Generating Symbolic Reasoning Problems with Transformer GANs"

Generating Symbolic Reasoning Problems with Transformer GANs This is the implementation of the paper Generating Symbolic Reasoning Problems with Trans

Reactive Systems Group 1 Apr 18, 2022
DTCN IJCAI - Sequential prediction learning framework and algorithm

DTCN This is the implementation of our paper "Sequential Prediction of Social Me

Bobby 2 Jan 24, 2022
Bottleneck Transformers for Visual Recognition

Bottleneck Transformers for Visual Recognition Experiments Model Params (M) Acc (%) ResNet50 baseline (ref) 23.5M 93.62 BoTNet-50 18.8M 95.11% BoTNet-

Myeongjun Kim 236 Jan 03, 2023
HINet: Half Instance Normalization Network for Image Restoration

HINet: Half Instance Normalization Network for Image Restoration Liangyu Chen, Xin Lu, Jie Zhang, Xiaojie Chu, Chengpeng Chen Paper: https://arxiv.org

303 Dec 31, 2022
A pytorch &keras implementation and demo of Fastformer.

Fastformer Notes from the authors Pytorch/Keras implementation of Fastformer. The keras version only includes the core fastformer attention part. The

153 Dec 28, 2022
[NeurIPS2021] Code Release of K-Net: Towards Unified Image Segmentation

K-Net: Towards Unified Image Segmentation Introduction This is an official release of the paper K-Net:Towards Unified Image Segmentation. K-Net will a

Wenwei Zhang 423 Jan 02, 2023
This project aims to segment 4 common retinal lesions from Fundus Images.

This project aims to segment 4 common retinal lesions from Fundus Images.

Husam Nujaim 1 Oct 10, 2021
Image reconstruction done with untrained neural networks.

PyTorch Deep Image Prior An implementation of image reconstruction methods from Deep Image Prior (Ulyanov et al., 2017) in PyTorch. The point of the p

Atiyo Ghosh 192 Nov 30, 2022
Pytorch implementation of MixNMatch

MixNMatch: Multifactor Disentanglement and Encoding for Conditional Image Generation [Paper] Yuheng Li, Krishna Kumar Singh, Utkarsh Ojha, Yong Jae Le

910 Dec 30, 2022
The versatile ocean simulator, in pure Python, powered by JAX.

Veros is the versatile ocean simulator -- it aims to be a powerful tool that makes high-performance ocean modeling approachable and fun. Because Veros

TeamOcean 245 Dec 20, 2022