
Not sure what would be the best way to find the right balance other than just try things randomly until something works. The first informational poster here, Influenza Frequently Complicated with Pneumonia, 1918, uses large text to emphasize words meant to catch the viewer’s eye, influenza and pneumonia, while the remaining text provides instructions to keep Chicago the healthiest city in the world. I guess this is where being a wizard in the black art of deep learning model tuning and debugging becomes handy.
COMBINING PERSONAL IMAGE PLUS TEXT FOR POSTER GENERATOR
Tried several combination of dropout, weight decay, but I could not seem to find the right balance. It takes no time to combine images using our online tool, just upload the images, and within seconds you will get the result. To make a poster that has to be printed larger than 56 inches, the working document has to be created at half the size of the final poster. Posters Flyers Brochures Social media Websites Stickers Yard signs QR Code Generator Browse Canva templates Choose from thousands of free, ready-to-use templates. After 2-3 epochs, the validation loss always became worst than the training loss.

In signage & wayfinding design color is the combining factor to harmonize the. My main problem right now is that my model overfitted almost immediately. If coloured text is used on a bright background the contrast will be weak. I saw that a fairly recent kaggle competition winner used a similar network architecture to mine: I wanted to explore if we could have better performance by putting everything in one neural network and let it figure it out.


Thank you for your input! I am still learning data science in general (specially new to kaggle competitions), but I saw that pretty much everyone did what you mentioned, mainly extract features from a CNN, extract features from an RNN for descriptions, then use those with the structured data in a boosting algorithm.
