Enhancing Model Accuracy through Advanced Transfer Learning Technique

[DarshAgrawal14]

Issue : https://github.com/Akasxh/Terrain-Recognition/issues/8

• The current implementation utilizes EfficientNet for terrain classification.
• Data Agumentation includes resizing, horizontal flipping, color jitter, and random rotations.
• I have loaded a pre-trained EfficientNet B0 model and modified the classifier layer to match the number of output classes (5 for different terrain types).
• I noticed at high learning rates the accuracy was highly volatile so it was lowered
• I defined the loss function (CrossEntropyLoss) and optimizer (Adam) with a lower learning rate for fine-tuning.
• Learning rate scheduling : ReduceLROnPlateau
• Included code to load model and test it using a example image
• The hyper - parameter tuning was performed using gpu , but if not available code will automatically use cpu
• Included code to load model and test it using a example image
• The hyper - parameter tuning was performed using gpu , but if not available code will automatically use cpu
• added tranfer learning notebooks , model , Readme.md ,improvement.md in transfer_learning directory
• Updated main README.md

Darsh Agrawal , GSSoC 2024 extd contributor

[DarshAgrawal14]

@Akasxh it seems the corrects labels are hactoberfest , gssoc-ext , level 3 , can you please update the labels so that the associated points are awarded.

[AKSHITHA-CHILUKA]

@Akasxh despite of many times telling you have not changed the labels. The crt labels are level3 with no spaces , gssoc-ext is the label not gssoc24-extd , hactoberfest-accepted is the label not hactoberfest . I hope you will change all the label's in all the issues and pr's as soon as possible .

[Akasxh]

@AKSHITHA-CHILUKA I am not available 24/7, I have updated the labels as soon I got back to my machine.

[AKSHITHA-CHILUKA]

so now you are available so I will see now whether you have added all the required labels or not

[Akasxh]

I have updated all the labels please check if there's any issue.

[AKSHITHA-CHILUKA]

All the issues and pr should be labeled

[Akasxh]

Should I be assigning once it's completed or just as a issue is raised?

The points are taken from prs right?

[AKSHITHA-CHILUKA]

both issue and pr if and only if both are labeled , then only the points will be reflected ,once you assign them you have to add all the labels including level labels .

[Akasxh]

Okay done

[AKSHITHA-CHILUKA]

ok for further doubts please connect with @Meetjain1