Prediction on fine_tunned model of 07_food_vision_milestone_project_1 is very poor.

[doffn]

I used your model and finetunned it:

then i made a prediction from the test data:

At the end the confusion matrix is really bad:

[MkRakesh]

Even I am facing issue with this model . I am getting 75% accuracy while evaluating, but while predicting the model prediction is always a specific class like it is predicting pizza as chicken curry :(

[doffn]

Can u share me your colab notebook. I donot want to pass to the next section with out fixing this.

[MkRakesh]

MkRakesh/Copy_of_Project_Food101.ipynb at main · MkRakesh/MkRakesh (github.com) <https://github.com/MkRakesh/MkRakesh/blob/main/Copy_of_Project_Food101.ipynb> I did try tuning different combination of layers like first 10 ,last10 then finally did all the EfficientNet layers.. please do have a look into the code. Excuse we if I have done something wrong with the code

…

On Thu, Aug 31, 2023 at 3:52 AM doffn ***@***.***> wrote: Can u share me your colab notebook. I donot want to pass to the next section with out fixing this. — Reply to this email directly, view it on GitHub <#579 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BAWUYXJ6XWC6E4D4MKJLUX3XX64JDANCNFSM6AAAAAA37AOJNA> . You are receiving this because you commented.Message ID: <mrdbourke/tensorflow-deep-learning/repo-discussions/579/comments/6869425@ github.com>