Using Pytorch, keras, Tensorflow and FastAI for implementing neural style transfer on given image dataset.
Using pretrained VGG19 model on pytorch neural style transfer was possible
- We first initialized the runtime on colab to GPU for faster processing
- Load the torch libraries and pretrained VGG19 model
- Get a content and a style image after defining the function to load the Image using PIL.Image
- Convert images to tensors
- Gram matrix method is done on the images
- training and validation steps are computed.
References:
##TODO 1)Keras 2) Tensorflow 3) fastAI
feel free to use and modify my repo and do let me know if any changes could be made. Cheers!!