Skip to content

carlosvint/ComputerVisionFinal

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 

Repository files navigation

ComputerVisionFinal

Final Project for Deep Learning Application II: Computer Vision

The repository contains the code to train a Conformer in a supervised manner and pre-train Conformer using SimCLR and fine-tune it on any dataset. The code also includes scripts to visualize the feature maps for conformer and perform a T-SNE analysis. Finally, I also include the code to train a Conformer using DINO.

Link to Conformer Paper and repository

Link to SimCLR Paper and repository

Link to DINO Paper and repository

For SimCLR pretraining in the SimCLR directory execute

python main.py --dataset CIFAR10

the model will be saved in the directory specified at config/config.yaml

To FineTune Conformer using a SimCLR pretrained model, specify the model path at the run.sh file and then run

sh run.sh

To visualize the feature map and attention layer output use:

python feature_maps.py -model_path model_path

About

Final Project for Deep Learning Application II: Computer Vision

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published