Machine Learning & Art

AI-generated Pokemon

This notebook allows you to create AI art in 2 clicks

What if AI had a more significant role in the artistic process of human beings? Would the goal of creativity be different if it were understood that artificial intelligence could create an object that was of value to sell?

The art industry will be at the mercy of AI technology The art industry is a global business that generates $200-billion in international sales. The growth of AI has led to many concerns that it will become the new way of doing business in the art industry.

A few examples of how AI is already impacting art are: The initiative by MLearning.ai worked with their client, Foundation St Anna, to create a digital version of a select piece from the Foundation’s collection. Using the power of AI technology, they created a digital version that was very similar in aesthetics to the original work.

This approach ensures that people can experience art in new ways and realize how far AI can take us into the future.

• April 2022 — AI art tools update can be found ➡️ HERE ⬅️

Many questions have arisen regarding how AI will impact art in our culture. Of course, we know that there is room for both things, but what’s important is to understand the value of art while introducing technology when it truly makes sense.

An example from this article: create a pokemon with two clicks, the creative process is kept to a minimum. The artist becomes an AI curator. Have fun!

#@title Download Models and Install/Load Packages (may take a few minutes)

!pip install rudalle==0.4.0 > /dev/null

from rudalle.pipelines import generate_images, show

from rudalle import get_rudalle_model, get_tokenizer, get_vae

from rudalle.utils import seed_everything

from huggingface_hub import hf_hub_url, cached_download

import os

import torch

import shutil

from IPython.display import clear_output

model_filename = "pytorch_model.bin"

device = "cuda"

config_file_url = hf_hub_url(repo_id="minimaxir/ai-generated-pokemon-rudalle", filename=model_filename)

cached_download(config_file_url, cache_dir='/content', force_filename=model_filename)

model = get_rudalle_model('Malevich', pretrained=False, fp16=True, device=device)

model.load_state_dict(torch.load(model_filename, map_location='cpu'))

vae = get_vae().to(device)

tokenizer = get_tokenizer()

!mkdir /content/gen_images > /dev/null

This notebook allows you to create AI-generated Pokemon in two clicks using a ruDALL-E model finetuned on the original images of all the Pokemon! This model was used to create Pokémon that resulted in AI-Generated Pokémon that went viral (10k+ retweets on Twitter + 30k+ upvotes on Reddit)

by Max Woolf

please don’t use this notebook to create NFTs k thx

To get started:

Copy this notebook to your Google Drive to keep it and save your changes. (File -> Save a Copy in Drive)

Run the cells below by clicking the Play button on the left of the cell (also visible when mousing-over the cell)

Art is everywhere 🟣 But the question is, how much do you love it?

I am an Art Curator, founder at EvArtology, and ML consultant at MLearning.ai. I advise companies and institutions in the creative industries on using AI tools in their daily work. Human collaboration with ML models can be very creative and bring huge benefits. The new era begins now.

• Twitter * Instagram * Tumblr * Linkedin * Facebook * AIcafe