I am a teacher as well and I teach programming at a local college in Canada (Originally from India). You are doing an excellent job and what is different about your teaching and other people is that you can make the material seem less daunting and scary. It gives us confidence that this can be learned with some work. Hopefully you will do more projects like this (maybe with financial data or fraud detection?)
Your teaching style is amazing sir. I had no prior knowledge of Deep learning or ML but I am able to follow this playlist. You teaching makes it too easy. Thank you!
So the goal of this project is to let farmers identify diseased leaves, right ? But when the farmers take a picture of the leaves, cant the farmers already tell if its diseased or not ?
Good question. Many times just by looking at it, it is not clear if it has a disease. That's when this application comes to a rescue. Also imagine an automated system where there is a big farm with cameras and these cameras are taking daily pictures and identifying the disease automatically. This saves time for a farmer as they don't have to check things manually.
Sir can u please show how the leaf disease detection can be predicted using advanced CNN. And how features can be extracted using optimization algorithm
Please help, I have error with cache: '''AttributeError Traceback (most recent call last) AttributeError: module 'tensorflow._api.v2.data' has no attribute 'AUTOTUNE''''
I have a concern about this code. Whenever the built in function 'take' is called to take a portion of the tensorflow dataset, the images and batches of the tensorflow dataset are reordered randomly. That means after making training dataset using first 80% images, when we take last 20% images for the test dataset some images of first 80% images can come in the last 20% images. This can cause same images in both the train and test dataset. Please clear my concern. Anyway, I have been watching your Deep Learning playlist from the first video. I have learnt many things from you. Thank you.
plt.figure(figsize=(10, 10)) for image_batch,label_batch in dataset.take(1): for i in range(12): ax= plt.subplot(3,4,i+1) plt.imshow(image_batch[0].numpy().astype("uint8")) plt.title(class_names[label_batch[0]]) plt.axis("off") its showing error 'tuple' object is not callable
Dear sir thank you so much for the video. Can you please post the videos on the implementation of compact convolution transformer and convolution vision transformer for potato disease detection sir
InvalidArgumentError Traceback (most recent call last) Input In [26], in ----> 1 history = model.fit( 2 train_ds, 3 batch_size=BATCH_SIZE, 4 validation_data=val_ds, 5 verbose=1, 6 epochs=EPOCHS, 7 ) help me out sir
Thank you for the awesome video..:) Can we also perform data augmentation using ImageDataGenerator class? it will be great if you can brief the difference between these two approaches(ImageDataGenerator and the .RandomFlip() etc. methods) of DA.
Yes you can, I think both methods do the same thing. I can see that ImageDataGenerator might be little easier, I will update my github notebook with that
@@codebasics do we need to use flow_from_directory or flo_from_dataframe, neither of them works on prefetchDataset, could you please include how to do DA using Imagedatagenerator for this .
i wonder if it would be a good filter to just run SVD on the images, may be some singular values and vectors would be identified that would easily split the images. Very curious.
Sir, I have a question: I am doing a research on coffee bean grading based on defect counts. The actual grading is done manually using a sample of 300gm of coffee bean. Based on defect occurrence grade is assigned. There might be multiple defects of same/different category in a given grade. The same defect type can be found in multiple grades. In this case how could I prepare a dataset? could I use the entire 300gm bean? Can I do both tasks; defect detection and classification? I would be grateful if you help me!
Hello. Your video is wonderful and i can learn alot here. But i am facing issue while using the same code from begning itself. Could you plz help me if possible.
How can we find out that each class of training testing and validation of the dataset for healthy , early Blight and late blight please tell me the solution I have confused. Dear sir kindly guide me
but why to manually split the dataset into 'training' and 'validation' set with the dataset.skip(54) function when you can simply set the parameter validation_split =0.8 inside the tf.keras.utils.image_dataset_from_directory module?
thanks for video, I have a question, why shuffle the train, test and validation data twice? in the function and then in this code line: train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size = tf.data.AUTOTUNE). thks
Hello Sir,Big Thanks for the great efforts.I am using google colab and i am unable to load the data.Can you please help,it is showing file not found error.
I am a MSc student for Information technology now , I interest to work final theses for deep leering in CNN for small image classification so please recommended for me to work these area which algorithm is better than the other ....please help me?
what were the libraries and its versions used here? because whenever i run data augmentation, it gives me a warning that randomflip and rotation does not have a converter. i used 3.9 python.
Hello, just saw this video, which is awesome by the way, but a question raised into my head. What about class balancing? I mean you have way fewer healthy pictures in your dataset. Isn't this a problem? how are you going to address this and how this impact the performance of your model? Thanks
Hello Sir, I am using google colab for this project and I am facing issue with importing the directory of PlantVillage, facing the error of cant find the directory, multiple times I checked the file path please provide a solution.
I also want to build this project from scratch i wanna know if i have to purchase any credentials for cloud services or this project doesn't include any payment for any services... pls reply anyone who has done the project
any one else getting error Found 0 files belonging to 1 classes. pasted in same directory stillgetting this error .ValueError: No images found in directory PlantVillage. Allowed formats: ('.bmp', '.gif', '.jpeg', '.jpg', '.png')