85b - An introduction to autoencoders - in Python

Thank you for this tutorial and a clear explanation. The only thing that left me wondering is the use of "accuracy" metric. Based on the mse loss (which is correct in this case), we're solving a regression task here, i.e. reconstructing continuous pixel values. In this case using MAE or RMSE for a metric seems more appropriate. Or did you choose the accuracy metric to measure the percentage of pixels that have been perfectly reconstructed by your autoencoder?

❤❤❤❤❤

do we need to normalize images if we do transfer learning or preprocessing functions of pretrained model will do it??

what does the model optimize for? the kernel matrix？

Your explanation is amazing. In simple words you explain the tough concepts. Thank you

Hello Sreenivas Your videos are extremely helpful and the content is presented very intuitively.
I had a thing to ask, what GPU enabled laptop would be the best to buy in your estimate? (I am not looking for specific names, but any pointers like the GPU specs. for eg the VRAM (8GB, 12GB or whatever) and # of cores that'll work decently for a few tasks at hand would be helpful. (Something a lil more exploitable than Google Colan that is). Thanks

keras and tensorflow versions ???

Excellent video !

Hey! You are doing a great work sir!
Can you make a video on installation of tf- gpu plz?

What is the metric "accuracy" mean in this task?