Project: Automatic labelling of Radiology Images

Ohh okay thanks @r0bby , Well i managed to push it on Gitlab using lfs. I’ll share both, :sweat_smile:

You did something very wrong. You don’t push your image to git, you push your Dockerfile to git. Read up on Docker https://docs.docker.com/get-started/

Okay, so I will upload the image on Docker Hub, and just share the docker file on the github Repository. Does that work?

Yes. That’s how you do it. I have no idea what you did to be honest. I also find it problematic that you don’t know how to use the git command-line – this is very important to Open Source.

I am using the git command line all this time, For uploading the file I thought you needed Docker Image file. So there’s an extension in Git which allows us to store the files efficiently on the repository. Later I realized I should’ve just sent the docker file and not the image.

There is no Docker Image file – we need the Dockerfile – you should read up on Docker because you’re likely doing a lot of things wrong.

1 Like

Hello @pri2si17, As requested, I have updated the Docker file in the repository. Hope that you are able to deploy it in your local machine. Do let me know if there’s any problem.

Thank you

Link the repo please?

Okay,

create a .dockerignore file with the following

__pycache__
*.pyc 

This will cut your image size down a lot.

Create a similar file called .gitignore with the following:

__pycache__
*.pyc 

Add and commit both.

Then do git rm __pycache__ *.pyc and commit.

This will cut both your repo and your docker image down a lot.

1 Like

Hello @pri2si17, I hope you got what you asked for. May I ask the future steps so that I can design a proper proposal? Also, you mentioned the requirement for main project will be updated. Will it be updated before the submission or after? Please do give your valuable feedback on the project :

Rename .dockerignore and .gitignore to remove the .txt extension

I just made some PRs to address this:

Thank you @r0bby, merged into the master

Also, you should have merged the dependabot PRs – those are typically security-related

Yes, I am working on it. The thing is that dependabot PRs update some file versions in the requirements.txt. When it is updated then the project crashes because of the tensorflow version of 1.15. I will be migrating the project to support tensorflow 2.0. Once that is done, I’ll update the requirements.txt accordingly.

Alrighty, wasn’t sure. Carry on. Dockerfile looks good. Your image definately shouldn’t be 2GB now lol

1 Like

Totally agree, Thank you very much :sweat_smile: @r0bby

It’s okay – I learned Docker over a couple years – enough to know best practices and to be able to glance at Dockerfiles and know when people are doing things wrong

1 Like

Thanks @r0bby for handling this :slight_smile: . @Parag0506 I would suggest you to not hurry, think before what you are doing and why you are doing? Wait till I review your code, but till then research existing models, their pros and cons and which model would be best and how to make your application user intuitive. This is a research based project, you need to do proper research.

1 Like

@SinghKislay any updates?