Image Classification using CNN from Scratch in Pytorch- Part 1 Training 

This explained the CNN model and the coding so well. I have been trying to find a video like this for ages. Thank you so much!!!!!!

after learning pytorch this was the only video with practical implementation .. Great video man

A bug that I couldn't find the whole night got resolved after watching this video. Thanks a ton!!

This is the most detailed and clear explanation one could have asked for.

Thank you so much for a detailed run-through on the shape of each layer!

If you're a newbie like me, practicing on a low-end Graphic card, following this video exactly, but found: "RuntimeError: CUDA error: out of memory” Reduce the batch_size in your train_load and test_loade. I downsize the batch_size to 64 to get it run without error.

Very very well explained.. Please keep up doing such great contents.

Best explanation ever, almost like cheating lol, given my months of studying this stuff

Thank you so much for this tutorial. Ive been looking for something like this for ages. Please do more videos like this. Kindly add automatic subtitles. Could you please do a tutorial on video object detection using pytorch?

awesome video man! you really helped me with my struggle with pytorch, thank you so much!

Interesting, really understand convolution

Thanks for making this amazing video, it would be better if you add the automatic subtitle.

Thank you so much for the video. This helped me a lot

You are best. It is really helpful. Thanks brother.

good！you deserve 1kk subscribers！

Thank you so much.💥

I was looking exactly for this , thanks much man :)

Awesome explanation

permission to learn sir. thanks

Excellent video

Very well explained

How image size Reduce to the factor of 2 with kernel_size=2 in max-pooling where stride and padding are not defined?

Hi, thank you for this awesome video. I would like to ask if I can run this model on the RPI4 and have my interference through the RPI?

very informative video. can you do another video for image classification using historical weather(LSTM) data and images (CNN).one neural network with two branches.

Amazing explaination. could you please make for RNNs, LSTMs, GANs , VAEs and Transformers ?

cuda is not found in Ge Force RTX 3080 NVIDIA graphics. How to fix it? I am using Pycharm for coding.

How did he know what kind of network with how many layers, kernels, strides etc he has to create? I mean I have been told that its trial based, but there should be some basis or foundation.

Thank you for this great tutorial

Great video!

On what basis we should select the no of output channels

Thanks really helpful

Please Help!. I don't understand why did you used 12 as output channels?

Hi and thank you for your video, how can I apply your tutorial to my dataset? I have already annotated the dataset and split them to train test and valid, I want to use MobileNet classifier for the task. Can you guide me?

my model accuracy is very less, its nearly 7% after implementing cnn model. what might be the reason

Can u send the data set for silkworm cocoon whether it is defective or not or can u explain it

Amazing Tutorial, this code is good for classification type task. I have project to perform semantic segmentation on images using CNN. Images contain "cancer image", and "Ground Truth" images. Please let me know how i perform this task using Pytorch? Finally, I want to calculate the IoU and Accuracy of proposed model.

can you please help me to evaluate this dataset using confusion matrix

Informative !

This is very helpful to me. I got a problem that " _init_() got an unexpected argument 'num classes' ". How can I fix this problem?

the test and train images, can they be the same or they have to be different

i am getting an error near " self.conv1=nn.Conv2d(in_channels=3,out_channel=12,kernel_size,stride=1,padding=1) " positional argument follows keyword argument (syntax error), can someone pls tell how to fix that

nice 1

Hello, can anyone tell me how to add the validation loss in here?

Thanks a lot

please make more lectures

why am i getting this error " in __iter__ raise TypeError('iteration over a 0-d tensor') TypeError: iteration over a 0-d tensor"

what is the architecture of model used here? architecture as in resnet, resnext, efiicientnet and so on..

sir, you done same method using another algorithm?

It may be a bit late but why do you use int as test_accuracy += int(torch.sum(prediction == labels.data)) I mean, every data we collect here is float always. Losses are float, so is accuracy. Why did we use 'int' there? Great tutorial by the way.

how to calculate f1 score ?

Is test folder have images

how much it take time to execute via cpu?

How can I split the own imagedataset into x train n y train

How the output channels is given as 12?

NameError: name 'ConvNet' is not defined how to fix??

hello why is there such a big difference between train acc and test acc?

🦈

bro forgot shift tab was a thing 23:21

Your accent is pretty weird, but I like your course hahah, thanks for your contribution