W2-Exploring Colab
Model 1: Neural Style TF
Find the model here.
I combined an image of Mr.Doodle's drawing with my photo of the Church in Iceland.
Model 2: Deepdream
Find the model here.
I think I like this version better than the one created with Runway. I felt I had more control of what’s happening (but also not really knowing what’s happening lol)
And this one is just really out of control. I cannot tell any connection with the original image.
Model 3: DeOldify-Image Colorizer
Find the model here
For this one, I wanted to test how good the coloring is with this model. So I picked an everyday object, blueberry, thinking the machine will easily recognize them. I turned the image into black and white in Photoshop and imported it into the model. The result was disappointing.
Results:
Model 4: Big Sleep
Find the model here.
“A Red Dog Standing on A Burger”
There were close to 30 steps, maybe more. These are the first, the last and some middle steps.
“Vegetable Pizza”
There were close to 40 steps, maybe more. These are the first, the last and some middle steps.
Model 5: 3D Photo Inpainting
Find the model here.