search

abhisheks008 / DL-Simplified

Deep Learning Simplified is an Open-source repository, containing beginner to advance level deep learning projects for the contributors, who are willing to start their journey in Deep Learning. Devfolio URL, https://devfolio.co/projects/deep-learning-simplified-f013
https://quine.sh/repo/abhisheks008-DL-Simplified-499023976
MIT License
360 stars 300 forks source link

Chocolate Classification using DL #572

Closed Arihant-Bhandari closed 4 months ago

Arihant-Bhandari commented 4 months ago

Pull Request for DL-Simplified 💡

Issue Title : #553 [Model Enhancement]: Chocolate Classification using DL

Closes: #553 [Model Enhancement]: Chocolate Classification using DL

Describe the add-ons or changes you've made 📃

introduced a new CNN architecture incorporating NADAM and ELU as optimizer and activation respectively, attention mechanism on overall model, changed data loading mechanism, introduced true image-extension checking allowing removal of unwanted images based on extensions like .webp, increased metrics. To prevent overfitting L2 regularizer or Ridge Regression is introduced to the model.

Type of change ☑️

What sort of change have you made:

How Has This Been Tested? ⚙️

Following metrics were used to evaluate the performance of the model:

Testing metrics on model:

Accuracy: 0.9696969696969697 Precision: 0.9090909090909091 Recall: 1.0 F1 Score: 0.9523809523809523 Confusion Matrix: [[22 1] [ 0 10]] AUC: 0.9782608695652174

Testing Loss: 0.3878534436225891

Checklist: ☑️

github-actions[bot] commented 4 months ago

Our team will soon review your PR. Thanks @Arihant-Bhandari :)

Arihant-Bhandari commented 4 months ago

@abhisheks008 hi, i have started the PR, pls check codebase, if any issues, discrepancies or errors, i would be more than glad to correct. if any suggestions regarding commenting and all, i would be glad to hear them. Thank you for your time reviewing.

Arihant-Bhandari commented 4 months ago

@abhisheks008 hi, as told, i have updated things into the original file, updated README with metrics and image, as well as enhanced readability of the README, mentioned enhancers with name and GITHUB-ID. the model, the .ipynb file and README have been sent as part of evaluation, pls check. thank you for your time and patience.

the existing file was untouched before previous submission, i have added things into the original as well as included comments as mush as possible for readability.

Arihant-Bhandari commented 4 months ago

ahh finally got what you talked abt on editing OG file, so sorry for this i guess i must have set outputs back to nill, i will correct this followup , extremely sorry for this mixup.

abhisheks008 commented 4 months ago

ahh finally got what you talked abt on editing OG file, so sorry for this i guess i must have set outputs back to nill, i will correct this followup , extremely sorry for this mixup.

No problem.

Arihant-Bhandari commented 4 months ago

hi @abhisheks008 as told, re-did things, re-did README, enhanced README, added acc image, added saved model. file is the original with my work as well as graph code for just plotting accuracy. i also added signatures with ID's for both readme and model enhancement.

also i wanted to ask if you could edit the README on both this as well as Colour Detection using DL, the attached images have not been configured to display CNN Attention acc graph, pls edit both README.md files.

thank you for your time and patience.

abhisheks008 commented 4 months ago

hi @abhisheks008 as told, re-did things, re-did README, enhanced README, added acc image, added saved model. file is the original with my work as well as graph code for just poltting accuracy. i also added signatures with ID's for both readme and model enhancement.

also i wanted to ask if you could edit the README on both this as well as Colour Detection using DL, the attached images have not been configured to display CNN Attention acc graph, pls edit both README.md files.

thank you for your time and patience.

If you wish you can do the changes, this will help you get an upgradation in the level of the issue. It's an opportunity for you to make this contribution as Level 3. @Arihant-Bhandari

Arihant-Bhandari commented 4 months ago

hi @abhisheks008 as told, re-did things, re-did README, enhanced README, added acc image, added saved model. file is the original with my work as well as graph code for just poltting accuracy. i also added signatures with ID's for both readme and model enhancement. also i wanted to ask if you could edit the README on both this as well as Colour Detection using DL, the attached images have not been configured to display CNN Attention acc graph, pls edit both README.md files. thank you for your time and patience.

If you wish you can do the changes, this will help you get an upgradation in the level of the issue. It's an opportunity for you to make this contribution as Level 3. @Arihant-Bhandari

hi so the thing is, when i originally pushed the file, the link in the Colour Detection README was for the bin in my own repo, on pushing the link changed to DL-Simplified bin, so if i want to add changes to that i would have to be issuing a README Enhancement issue for it. similarly for Chocolate detection, thats the issue with image you have tagged as change. my code file is ready i have made all the necessary changes for this issue's completion.

hence i would like to know your opinion on the grading scheme for these PRs, would you allow another issue to be raised for both project's README enhancements for images at same time in single issue ?

thanks for listening

Arihant-Bhandari commented 4 months ago

the image bin link is not updated dynamically on push to this repo's image directory, so if you go back to Colour Detection using DL folder, the README would have CNN-Attention image missing since the link is set on my cloned repo, which is not getting picked due to different repos.

Arihant-Bhandari commented 4 months ago

i was checking up on things in the repo and noticed this, hence leaving the Chocolate classification README with empty link so that a new issue can be raised and committed to setting link properly

Arihant-Bhandari commented 4 months ago

would you accept raising this issue to level3 if i were to say, update that README in this PR ? so this PR would finally include edited .ipynb file with CNN-ATTENTION, model for this, readme file with missing picture link whcih i will update in the next proeject i had asked you for , and finally the image file for this project.

abhisheks008 commented 4 months ago

If you can do the changes in this PR itself then only it can be raised to Level 3. If you are planning to enhance the READMEs by creating separate issues then, you'll get 1 point each for every README enhancement. Project admins had a clear discussion with the core team regarding the README enhancment issues, as many are using generative scripts and making changes in the README. That's why I am not accepting any README enhancement issues from the contributors.

Arihant-Bhandari commented 4 months ago

Ohhhh , cool then , I will edit the colour detection README , push in this PR and let you know , thanks for replying @abhisheks008

Arihant-Bhandari commented 4 months ago

@abhisheks008 i have edited the files as required, kindly review and approve the PR, thank you for your time and patience in guiding me ❤️.

Arihant-Bhandari commented 4 months ago

@abhisheks008 i have changed the README so that it can pick up the image from your end as soon as this PR is merged, i just configured the paths pointing towards your file which is currently nonexistent but will start existing on PR merge. the image is one of the commits i am pushing in this PR.

abhisheks008 commented 4 months ago

image

Please fix this issue.

Hi @Arihant-Bhandari this issue is not fixed yet! :(

Arihant-Bhandari commented 4 months ago

Hi, the issue is coming up because the link is the future repo's link , the moment this PR is pushed , the link will be active to be displayed on the README

Arihant-Bhandari commented 4 months ago

If I were to add the link on my repo and push it will not be able to trace back the link to pushed files, that's why I have made things so that the moment this PR is pushed , things start working on the README.md end.

abhisheks008 commented 4 months ago

No need to put that, use the existing/current URL, GitHub will automatically fix that once that is merged in the main repo.

Arihant-Bhandari commented 4 months ago

@abhisheks008 hi , I have fixed all issues , pls check

Appropriate files have been added , README for colour detection has been corrected , README for chocolate detection has been enhanced , best model for CNN-Attention has been added as part of submission along side original code and ACCURACY graph for CNN-Attention

Arihant-Bhandari commented 4 months ago

thank you for your time and patience for this project, really glad to work on this