Python tips and tricks - 7: Continuing keras model training when using custom loss and metrics

Loading a keras model and continuing training​
When using custom loss function and metrics​.

No code to share with this video.
Summary: Provide your custom optimizer or loss or metrics as custom objects during loading the model.

custom_objects={'my_custom_loss': custom_loss,
'custom_metric': my_custom_metric})

[sharvildandekar]

Thank you for the entire ML and DL playlist. If all hope is lost and I cannot find a solution to my problem, I know it'll be there somewhere on your channel. Could you also make a video on Spatio-Temporal Predictions using GANS

[ashberten]

Thank you for your nice tutors. I would love if you could make a video on "Test Time Augmentation".

[internetexplorer]

Thank you for your tutorials. It really helped me a lot with my research 🙏

[IdanCs]

Usually when training, epochs tend to take longer time to compute as we progress. I found, however, that by training the model and then re-loading it and continue the training - the epochs take much shorter time to finish. Has anyone else noticed that?

[sowmyabhat]

Could you please explain the concept of albumentation in Unet segmentation of data ..

[RG-jcuq]

I have a question. I want calculate arousal and valence amotion from videos with deep learning in pyrhon. But i dont know how cab do this. Can you help me how can do this?
Thanks for your best videos ❤️❤️🙏🙏

[nagarajp.y]

Thank you for your nice tutors sir. Could you explain us the concepts of 2D to 3D reconstruction...