Train the Network with colour images instead of b/w

[Xx-Ashutosh-xX]

Train the Network with colour images instead of b/w i.e. use 3 channels in the CNN.

[pavelxx1]

hi,thx
but how to do correct labels for images?

[Xx-Ashutosh-xX]

labels remain the same because in the model in outputting 3 signals {buy, sell, hold}. I just wanted to test out if instead of training the model on black/white images we use colour images which might increase its accuracy.

See create_image.py if you want to know about labelling.