👉 Check out the blog post and other resources for this video: 🔗 deeplizard.com/learn/video/bfQBPNDy5EM 👀 Come say hey to us on OUR VLOG: 🔗 th-cam.com/users/deeplizardvlog
Thank you very much for this video! It was a good video for explaining how the prediction output can be understood in terms of which class is being predicted (i.e. whether "0" or "1") is being predicted!
Thanks i really like this simple and straightforward way of explaining the code without overwhelming the learner. On a side note i see almost all examples on CNN being cats and dogs, the ML community needs to switch to Penguins and parrots.
Thank you for creating these tutorials, I would like to point out that there is a shortage of tutorials on functional API compared to the sequential. If you can look into it that will be great
Are there any resources you could provide that actually tests the model? Id like to give the model a photo and have it print to the console whether it is a cat or dog. thanks
Why do we need to set the testing set to unshuffle in ImageGenerator? For my understanding, shuffled the testing set won't change the sample and its corresponding data so it won't influence the correctness or plotting. I tried to see the observe the difference by plotting the shuffle and unshuffle set and couldn't see the difference. Could you clarify more? Thank you!
Yes, I am very confused on this. I assume that when you shuffle, the corresponding labels shuffle in the same way as the samples. Such that each sample still maps to the correct label. If this was not the case, how would the CNN learn from the training data?? So shuffling the test data should have no affect on the mapping from samples to labels. Did you find the answer to this question? If so, please share it with me.
I understand now. The labels are seperate from the classes. The labels represent one-hot encodings (Usually used for training) and the classes represent binary classifications. 0 for cat and 1 for dog. When you shuffle the test data it does not shuffle the classes, it only preserves the mappings of the one-hot encoding labels. She uses the classes labels on the confusion matrix. An easy fix to not being able to shuffle is to get the argmax of the test_labels aka the one hot encodings. Reply to me if you want a better explanation.
Sir&Ma'am, I want to build a PC for Deep Learning purposes. My budget is 1.6L INR or 2135 USD. Please suggest me the best build for my budget. I am your fan from India.
You can follow the order of the Keras course by viewing the course on deeplizard.com or the TH-cam playlist for the course. Both linked below. deeplizard.com/learn/playlist/PLZbbT5o_s2xrwRnXk_yCPtnqqo4_u2YGL th-cam.com/play/PLZbbT5o_s2xrwRnXk_yCPtnqqo4_u2YGL.html
👉 Check out the blog post and other resources for this video:
🔗 deeplizard.com/learn/video/bfQBPNDy5EM
👀 Come say hey to us on OUR VLOG:
🔗 th-cam.com/users/deeplizardvlog
Your channel is very helpful. Thank you
thanks for the video , i really like the way u teach great work , keep going ....
Thank you very much for this video! It was a good video for explaining how the prediction output can be understood in terms of which class is being predicted (i.e. whether "0" or "1") is being predicted!
Thanks i really like this simple and straightforward way of explaining the code without overwhelming the learner.
On a side note i see almost all examples on CNN being cats and dogs, the ML community needs to switch to Penguins and parrots.
🐧🦜 ➡ next vid 😅
Hey Mandy !! Learned much of Machine Learning from your channel. Would love to see a video on Recurrent Neural Networks
It's perfect if you show some techniques to overcome that overfitting problem!
Anyway, thank you so much!!
Thank you for creating these tutorials, I would like to point out that there is a shortage of tutorials on functional API compared to the sequential. If you can look into it that will be great
Great. Thanks.
Are there any resources you could provide that actually tests the model? Id like to give the model a photo and have it print to the console whether it is a cat or dog. thanks
Thanks for the video. Guys, Would you please suggest a way for me to learn the Mathematics behind Deep Learning?
@E B T, go for Machind learning by Stanford University by Andrew NG
Thanks bro.
Hi.
Confusion Matrix has white text on white background, how can i solve?
How is the model overfitting? Didn't we plot the confusion matrix for the test set? Isn't overftting the case when model is too good on train set?
Hey!! Team, Can you make a video on diabetic retinopathy dataset and perform CNN algorithm on it
Why do we need to set the testing set to unshuffle in ImageGenerator? For my understanding, shuffled the testing set won't change the sample and its corresponding data so it won't influence the correctness or plotting. I tried to see the observe the difference by plotting the shuffle and unshuffle set and couldn't see the difference. Could you clarify more? Thank you!
Yes, I am very confused on this. I assume that when you shuffle, the corresponding labels shuffle in the same way as the samples. Such that each sample still maps to the correct label. If this was not the case, how would the CNN learn from the training data?? So shuffling the test data should have no affect on the mapping from samples to labels.
Did you find the answer to this question? If so, please share it with me.
I understand now. The labels are seperate from the classes. The labels represent one-hot encodings (Usually used for training) and the classes represent binary classifications. 0 for cat and 1 for dog. When you shuffle the test data it does not shuffle the classes, it only preserves the mappings of the one-hot encoding labels. She uses the classes labels on the confusion matrix.
An easy fix to not being able to shuffle is to get the argmax of the test_labels aka the one hot encodings. Reply to me if you want a better explanation.
Sir&Ma'am,
I want to build a PC for Deep Learning purposes. My budget is 1.6L INR or 2135 USD.
Please suggest me the best build for my budget.
I am your fan from India.
es bonita uu
hello sorry but your videos seem to be out of order. I've been trying to find the 'last couple of videos' that explain how to create the test set .
You can follow the order of the Keras course by viewing the course on deeplizard.com or the TH-cam playlist for the course. Both linked below.
deeplizard.com/learn/playlist/PLZbbT5o_s2xrwRnXk_yCPtnqqo4_u2YGL
th-cam.com/play/PLZbbT5o_s2xrwRnXk_yCPtnqqo4_u2YGL.html
@@deeplizard thank you. your videos are helping me with my uni project.