Oooh your final image / layered self portrait is really breathtaking! Yeah it takes some time to really get the hand of Deep Dream Generator. For one thing you have to think about how the AI will "see" a style image, as it is quite a lot different from how you or I would see it. Properly cropped, a section of picasso could work for what you are trying to do but in general it will not majorly distort the larger shapes in the base image - if you are after a distorted picasso output, I would play around with distorting the base image before you run it through Deep Dream generator, then using the generator to apply Picasso-esque detail to the brushstrokes and color palette. Also playing around with the settings as far as increasing or decreasing the scale of the style image, and weight of style image vs. base image, can result in huge changes in your final product. I have been playing around with it pretty much nonstop for a half a year now and finally feel like I am attaining a higher degree of proficiency, but even now it is still a lot of repeat attempts and trial and error to get results like I shared in my posting :) Much love - Carl