Python script to preprocess images of all Pokémon to finetune ruDALL-E

Overview

ai-generated-pokemon-rudalle

Python script to preprocess images of all Pokémon (the "official artwork" of each Pokémon via PokéAPI) into a format such that it can be used to finetune ruDALL-E using the finetuning example Colab Notebook linked in that repo. This workflow was used to create a model that resulted in AI-Generated Pokemon that went viral (10k+ retweets on Twitter + 30k+ upvotes on Reddit)

My modified Colab Notebook that I used to finetune the model on Pokémon is here: this Notebook's release is purely for demonstration/authentication purposes and no support will be given on how to use it because it is incredibly messy and embarrassing, but there may be a few ideas there that are useful for future generation. Some notes on how the process works are included below, with oppertunity to reproduce/improve it.

The script outputs two things: an images folder with all the preprocessed images plus a data_desc.csv file which contains the image path and Russian caption pairs for finetuning. Some examples of the preprocessed input images are present in the images folder, plus the final data_desc.csv.

The model used is not included in this repo because it's currently too large (~3GB) to distribute (will add the model to Hugging Face at some point).

Preprocessing Script Notes

  • The GraphQL interface to PokéAPI is used as it allows to retrieve the type information plus IDs of all Pokémon in a single request. As a bonus, the returned IDs include the alternate forms of Pokémon (e.g. Mega) which would not otherwise be present just by incrementing IDs.
  • ruDALL-E requires 256x256px, RGB input images. In this case the source input images from PokéAPI are conveiently both square and larger than 256x256 so they downsample nicely. Since the images have transparency (RGBA), they are composited onto a white background.
  • The translation service used is Yandex, which apparently has decent rate limits, plus as a Russian company the translations from English to Russian should theoetically be better.
  • The captions (which are later translated into Russian) are determined by type. For example, a Grass/Poison type will have the caption A Grass-type and Poison-type Pokémon, which is then translated into Russian. In theory, this improves the finetuning process by allowing ruDALL-E to notice trends, plus in theory this can be leveraged at generation-time to control the generation (e.g. prompt with A Grass-type Pokémon and have ruDALL-E generate only Grass-type Pokémon)
  • Due to potential rate limits on translation, translations are cached at runtime by Pokémon type(s) so the API is pinged only once.

Finetuning and Generation Notes

  • The model used above was trained for 12 epochs (4.5 hours on a P100), at a max learning rate of 1e-5. The pct_start param of the OneCycleLR scheudler was set to 0.1 so that learning rate decay happens faster. Despite that, the model converged quickly.

  • The parameters for finetuning ruDALL-E are very difficult to get the expected results. Too little training and the output images will be too incoherent; too much training and the model will overfit and output the source images, and also ignore any text prompts. In the social media posts above, the model is slightly overfit and attempts at using text prompts to control generation failed. But overfitting is not necessairly a bad thing as long as it avoids verbatim output.

Usage

You can install the dependences via:

pip3 install Pillow requests translatepy tqdm

Then run build_image_dataset.py

Getting the images into the ruDALL-E finetuning Colab Notebook is up to the user, but the recommended way to do so is to ZIP the generated images folder (~42 MB!), upload it to Colab (or upload to Google Drive and copy it into the Notebook from there), and unzip the folder in Colab itself via !unzip.

Maintainer/Creator

Max Woolf (@minimaxir)

Max's open-source projects are supported by his Patreon and GitHub Sponsors. If you found this project helpful, any monetary contributions to the Patreon are appreciated and will be put to good creative use.

License

MIT

Owner
Max Woolf
Data Scientist @buzzfeed. Plotter of pretty charts.
Max Woolf
Versión preliminar análisis general de Covid-19 en Colombia

Covid_Colombia_v09 Versión: Python 3.8.8 1/ La base de datos del Ministerio de Salud (Minsalud Colombia) está en https://www.datos.gov.co/Salud-y-Prot

Julián Gómez 1 Jan 30, 2022
My collection of mini-projects in various languages

Mini-Projects My collection of mini-projects in various languages About: This repository consists of a number of small projects. Most of these "mini-p

Siddhant Attavar 1 Jul 11, 2022
A simple script written using symbolic python that takes as input a desired metric and automatically calculates and outputs the Christoffel Pseudo-Tensor, Riemann Curvature Tensor, Ricci Tensor, Scalar Curvature and the Kretschmann Scalar

A simple script written using symbolic python that takes as input a desired metric and automatically calculates and outputs the Christoffel Pseudo-Tensor, Riemann Curvature Tensor, Ricci Tensor, Scal

2 Nov 27, 2021
Easy Alias's for bash

easy-alias Easy Alias's for bash Setup Your system needs to have 'echo' which every 21st century computer has You dont need any python requirments but

Hashm 2 Jan 18, 2022
Python communism - A module for initiating the communist revolution in each of our python modules

Python communist revolution A man once said to abolish the classes or something

758 Jan 03, 2023
Project5 Data processing system

Project5-Data-processing-system User just needed to copy both these file to a folder and open Project5.py using cmd or using any python ide. It is to

1 Nov 23, 2021
Integration of CCURE access control system with automation HVAC of a commercial building

API-CCURE-Automation-Quantity-Floor Integration of CCURE access control system with automation HVAC of a commercial building CCURE is an access contro

Alexandre Edson Silva Pereira 1 Nov 24, 2021
OpenSea NFT API App using Python and Streamlit

opensea-nft-api-tutorial OpenSea NFT API App using Python and Streamlit Tutorial Video Walkthrough https://www.youtube.com/watch?v=49SupvcFC1M Instruc

64 Oct 28, 2022
Pykeeb - A small Python script that prints out currently connected keyboards

pykeeb 🐍 ⌨️ A small Python script that detects and prints out currently connect

Jordan Duabe 1 May 08, 2022
Camera track the tip of a pen to use as a drawing tablet

cablet Camera track the tip of a pen to use as a drawing tablet Setup You will need: Writing utensil with a colored tip (preferably blue or green) Bac

14 Feb 20, 2022
Groupe du projet Python en 2TL2-4

Présentation Projet EpheCom Ce logiciel a été développé dans le cadre scolaire. EpheCom est un logiciel de communications - vocale et écrite - en temp

1 Dec 26, 2021
Ferramenta de monitoramento do risco de colapso no sistema de saúde em municípios brasileiros com a Covid-19.

FarolCovid 🚦 Ferramenta de monitoramento do risco de colapso no sistema de saúde em municípios brasileiros com a Covid-19. Monitoring tool & simulati

Impulso 49 Jul 10, 2022
Automation of VASP DFT workflows with ASE - application scripts

This repo contains a library that aims at automatizing some Density Functional Theory (DFT) workflows in VASP by using the ASE toolkit.

Frank Niessen 5 Sep 06, 2022
Custom Weapons 3 attribute support for Custom Weapons X

CW3toX Allows use of Custom Weapons 3 attributes in Custom Weapons X. Requiremen

2 Mar 01, 2022
Simple Assembler with python

Assembler with python converts assembly source code to machine code Requirements Python 3 🐍 Usage python main.py [source] [output] [source] : Path t

Amir mohammad 1 Dec 24, 2021
Python client library for the Databento API

Databento Python Library The Databento Python client library provides access to the Databento API for both live and historical data, from applications

Databento, Inc. 35 Dec 24, 2022
Your E-Canteen that is convenient and accessible wherever you are in the campus

Food Web E-Canteen System Your E-Canteen that is convenient and accessible wherever you are in the campus. Table of Contents About The Project Contrib

Pudding 5 Jan 07, 2023
In this project, we'll be creating a virtual personal assistant for ourselves using our favorite programming language

In this project, we'll be creating a virtual personal assistant for ourselves using our favorite programming language, Python. We can perform several offline as well as online operations using the bo

Ashutosh Krishna 188 Jan 03, 2023
Program Input Data Mahasiswa Oop

PROGRAM INPUT NILAI MAHASISWA MENGGUNAKAN OOP PENGERTIAN OOP object-oriented-programing/OOP adalah paradigma pemrograman berdasarkan konsep "objek", y

Maulana Reza Badrudin 1 Jan 05, 2022
MIT version of the PyMca XRF Toolkit

PyMca This is the MIT version of the PyMca XRF Toolkit. Please read the LICENSE file for details. Installation Ready-to-use packages are available for

V. Armando Solé 43 Nov 23, 2022