A question for me someone solve this in 1:35:41 he is using the same train_datagen variable for both testing and validation ..This is not the right way i think train_generator = train_datagen.flow_from_directory('tumorous_and_nontumorous/train/', batch_size=32, target_size=(224, 224)) test_generator = test_data_gen.flow_from_directory('tumorous_and_nontumorous/test/', batch_size=32, target_size=(224, 224)) valid_generator = valid_data_gen.flow_from_directory('tumorous_and_nontumorous/valid/', batch_size=32, target_size=(224, 224)) training data should have the augmentation the testing and validation data should not have augmentation this shows how the model performs in the real world am i wrong here?? please someone correct me if i am
I am getting this kind of error at @1:09:56. error: OpenCV(4.9.0) /Users/xperience/GHA-OpenCV-Python2/_work/opencv-python/opencv-python/opencv/modules/imgproc/src/color.cpp:196: error: (-215:Assertion failed) !_src.empty() in function 'cvtColor' Kindly provide me a solution for this .
@@knowledgedoctor3849 I am getting this error when I commented the resize line of code ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (3000,) + inhomogeneous part. Could you please be of help
@know Sorry, I want to ask where you originally got this vgg_unfrozen.h5 file from, because I watched the video and only saw you pasting this file without saying what it was or where it came from
Sorry, I want to ask where you originally got this vgg_unfrozen.h5 file from, because I watched the video and only saw you pasting this file without saying what it was or where it came from
bro you did a great job !! but i found that the f1 score = 0.5 and recall=0.49 wich is low can you fix this or help me fix this please i need this help
Bro ho Chuka hai Kam b Kar Raha but main mazeed data add krna Chahta hun Kesy kron or accuracy Kesy increase kron mein ne ephoc b 2 Sy 5 Kar deye but still accuracy less than 75% so please guide kren bro
@@knowledgedoctor3849 Bro Actually (vgg_unfrozen.h5 )File is Missing if You Don't Mind Can You Please Send me It by today in Instagram( user id: coz_im_proplayer) I sent You An Invite Tomorrow we Have Project Review Bro
@@knowledgedoctor3849 Bro Actually (vgg_unfrozen.h5 )File is Missing if You Don't Mind Can You Please update it in Git by today Tomorrow we Have Project Review Bro
Sir i have different dataset for a different project and following your video for that project , i have 2240 images in class 1 and 3200 images in class 2 in my dataset, i dont want to do data augmentation, can i proceed without data augmentation ? Or data augmentation is necessary ?
Sir after running model 2, i am getting testing accuracy of 72% But after running model 3, i am getting testing accuracy of 67% How to improve final testing accuracy?
I am getting this error when I commented the resize line of code ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (3000,) + inhomogeneous part.
am getting this error when I commented the resize line of code ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (3000,) + inhomogeneous part. image = crop_brain_tumor(image, plot=False) # image = cv2.resize(image, dsize=(image_width, image_height), interpolation = cv2.INTER_CUBIC) # image = image/255.00 X.append(image) if directory[-3:] == "yes":
Sorry, I want to ask where you originally got this vgg_unfrozen.h5 file from, because I watched the video and only saw you pasting this file without saying what it was or where it came from
@@vg1603 Sorry, I want to ask where you originally got this vgg_unfrozen.h5 file from, because I watched the video and only saw you pasting this file without saying what it was or where it came from
Great Work Brother and Keep Rocking 👏
You only use just 2 epochs to get a slow accuracy, Is that's correct to make the final classification?
A question for me someone solve this
in 1:35:41 he is using the same train_datagen variable for both testing and validation ..This is not the right way i think
train_generator = train_datagen.flow_from_directory('tumorous_and_nontumorous/train/', batch_size=32, target_size=(224, 224))
test_generator = test_data_gen.flow_from_directory('tumorous_and_nontumorous/test/', batch_size=32, target_size=(224, 224))
valid_generator = valid_data_gen.flow_from_directory('tumorous_and_nontumorous/valid/', batch_size=32, target_size=(224, 224))
training data should have the augmentation the testing and validation data should not have augmentation this shows how the model performs in the real world
am i wrong here?? please someone correct me if i am
I am getting this kind of error at @1:09:56.
error: OpenCV(4.9.0) /Users/xperience/GHA-OpenCV-Python2/_work/opencv-python/opencv-python/opencv/modules/imgproc/src/color.cpp:196: error: (-215:Assertion failed) !_src.empty() in function 'cvtColor'
Kindly provide me a solution for this .
did you get the solution
i am getting same error
broder , ponle en def crop_brain_tumor(image, plot=False): import cv2 , falta ponerle eso y funciona!
i had the same problem the problem is in this line:
folder1 = 'augmented_data/no'
the solution is adding a '/'
folder1 = 'augmented_data/no/'
Sir I am working on Knee Osteoarthritis project which make sure of DL Can you also do that project so that I can get some help from it.
Thank you for the video.
Can I use this for unfixed image size classification
Yeah sure
@@knowledgedoctor3849 I am getting this error when I commented the resize line of code
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (3000,) + inhomogeneous part.
Could you please be of help
You does not mentioned the Dataset source.
brilliant ! video plz continue like this
Good job !
but how can i know the type of tumor ?
aug images are not loading or showing me in my file...then what supposed to do?
@know Sorry, I want to ask where you originally got this vgg_unfrozen.h5 file from, because I watched the video and only saw you pasting this file without saying what it was or where it came from
Sorry, I want to ask where you originally got this vgg_unfrozen.h5 file from, because I watched the video and only saw you pasting this file without saying what it was or where it came from
He runs the file on google collab on 30 epochs and downloaded the unfrozen model
What happens when you upload a picture from the internet or a picture that was not used to train the AI, does it predict it well?
omg! nice question! did you try that?
bro you did a great job !! but i found that the f1 score = 0.5 and recall=0.49 wich is low can you fix this or help me fix this please i need this help
overfit??
How can I do this work in pycharm please
Yes You Can.
ModuleNotFoundError: No module named 'distutils'
how to solve this error ?
import the module
Bhai jab mein vs code me run krta hun code ko to muju cv2 module not found error de raha h kia kron
Aap ko Opencv Install Karna Padega Bhai, Then Kam Karega...
pip install opencv-python
pip install opencv-contrib-python..
Bro ho Chuka hai Kam b Kar Raha but main mazeed data add krna Chahta hun Kesy kron or accuracy Kesy increase kron mein ne ephoc b 2 Sy 5 Kar deye but still accuracy less than 75% so please guide kren bro
Bro Can You also Upload or Send The Files Generated Using Jupiter Notebook Because It Show Me The Files are missing
Ok Let Me Upload Clean Code With Graph
@@knowledgedoctor3849 Bro Actually (vgg_unfrozen.h5 )File is Missing if You Don't Mind Can You Please Send me It by today in Instagram( user id: coz_im_proplayer) I sent You An Invite
Tomorrow we Have Project Review Bro
@@knowledgedoctor3849 Bro Actually (vgg_unfrozen.h5 )File is Missing if You Don't Mind Can You Please update it in Git by today Tomorrow we Have Project Review Bro
Sir i have different dataset for a different project and following your video for that project , i have 2240 images in class 1 and 3200 images in class 2 in my dataset, i dont want to do data augmentation, can i proceed without data augmentation ? Or data augmentation is necessary ?
Seems like your dataset is balanced. Don't need to use data augmentation.But it will help you to build your model more accurate that's it..
@@knowledgedoctor3849ok sir, Thank You very much
Broo where i can find final "vgg_unfrozen" file
U found it ?
Below the GitHub link
Sir after running model 2, i am getting testing accuracy of 72%
But after running model 3, i am getting testing accuracy of 67%
How to improve final testing accuracy?
Check that that's you give here correct model or not... Cz It's little bit confusing..
Check that again
@@knowledgedoctor3849 sir I have given the correct model
For final testing the code is
Model_03.load_weights("model_weights/vgg_unfrozen.h5")
Sir the model is correct, I don't know why that is happening
Can you please help me@@knowledgedoctor3849
@@knowledgedoctor3849
SIr the model is correct, but it is still not working,
please help me sir.
Dataset ??pls provide the dataset
It is available on Kaggle
@@krishnaverma86 link pls
I am getting this error when I commented the resize line of code
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (3000,) + inhomogeneous part.
where can i find all the dataset please !
At Github Repo Please Check
am getting this error when I commented the resize line of code
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (3000,) + inhomogeneous part.
image = crop_brain_tumor(image, plot=False)
# image = cv2.resize(image, dsize=(image_width, image_height), interpolation = cv2.INTER_CUBIC)
# image = image/255.00
X.append(image)
if directory[-3:] == "yes":
bhai vgg_unforzen.h5 file kha h?
Let me upload it at drive🌻
@@knowledgedoctor3849 bhaiya mil nhi rha
@@anjalisinghal8959file mila kya apko?
Where is drive
Sorry, I want to ask where you originally got this vgg_unfrozen.h5 file from, because I watched the video and only saw you pasting this file without saying what it was or where it came from
Where is the file vgg_unfrozen.h5
Below the GitHub link
It's a detection problem, not a classification problem.
Hyy, did you find the complete detection+classification source code anywhere, if so, could you tell me about it
@@vg1603
Sorry, I want to ask where you originally got this vgg_unfrozen.h5 file from, because I watched the video and only saw you pasting this file without saying what it was or where it came from
Where is the file vgg_frozen.h5