search

Akasxh / Terrain-Recognition

High accuracy, explainable, lightweight CNN for terrain recognition.
GNU General Public License v3.0
11 stars 16 forks source link

Terrain Recognition using RESNET50 #17

Open ramana2074 opened 1 month ago

ramana2074 commented 1 month ago

Pretrained models have already learned to recognize general features and patterns in data, which can improve their performance on your specific task.

AKSHITHA-CHILUKA commented 1 month ago

@Akasxh despite of many times telling you have not changed the labels. Please add 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 commented 1 month ago

Could you please completely make a seperate folder with name of technique used just as transfer learning folder that has been made. Use similar architecture and make sure to include all those components in them.

ramana2074 commented 1 month ago

added into a residual learning file

Akasxh commented 1 month ago

Please add Plot training and validation accuracy over epochs. This is crucial to study how different techniques are working.

ramana2074 commented 1 month ago

added Plots

Akasxh commented 1 month ago

Try making your model more efficient from what I've see your accuracy is pretty low. image

ramana2074 commented 1 month ago

terrain output is this okay?

Akasxh commented 3 weeks ago

Sorry to say it's not yet fully optimised try adding some multimodality of resnets itself to see how they turn out to be.