Fruits and Vegetables Recognition System Part-4 | Training and Compiling Image Recognition Model
HTML-код
- Опубликовано: 11 сен 2024
- Hello and Welcome guys
In this project we'll learn how to make powerful deep learning model for 36 different classes of image
In this video,
We'll compile convolution neural network using tensorflow
We'll also perform training of convolution neural network and also understand how to choose different parameter values , what is the meaning of each parameter
Playlist Link: • Fruits and Vegetables ...
Project Link:
github.com/ani...
Dataset Link:
www.kaggle.com...
sir i understood that you used CNN architecture, but is there any specific CNN type you are using
No here whatever architecture is developed its developed from scratch using tensorflow
@@SPOTLESSTECH thank you sir
Error is displaying after train_function that unexpected result of train function
At 4:00 after runnign this command , i got a value error saying " ValueError: Arguments `target` and `output` must have the same shape. Received: target.shape=(None, 36), output.shape=(None, 7056) ". Please Help
Change number of units value in output layer equal to number of classes present in your dataset
If you are using same dataset as in video then change it to 36
@@SPOTLESSTECH I got my error fixed by a very strange thing suggested by gemini , and that was to use cnn.compile (this line) with training history line and run them together , i dont know what was the reason behind it, but worked for me.
@@SPOTLESSTECH Thanks for replying man , i appreciate it
Error is displaying after training_history=cnn that unexpected result of train function
Dm details with screenshot of error on Instagram page will help you to fix that
bro after compiling my accuracy is only 0.028 (very low) i'm confused pliz 🙏help!!!!!!!
Check your model architecture is it same as in video
i compiled again from the beginning and change the optimzer to 'adam' and works fine....thanks
error is displaying after training function, i've sent a dm to your instagram about it
Will check it out