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
Data derived from the OpenType specification

This package currently provides the opentypespec.tags module, which exports FEATURE_TAGS, SCRIPT_TAGS, LANGUAGE_TAGS and BASELINE_TAGS dictionaries, representing data from the Layout Tag Registry

Simon Cozens 4 Dec 01, 2022
Free components that wrap up Python into Delphi and Lazarus (FPC)

Python for Delphi (P4D) is a set of free components that wrap up the Python DLL into Delphi and Lazarus (FPC). They let you easily execute Python scri

747 Jan 02, 2023
Python Project Template

A low dependency and really simple to start project template for Python Projects.

Bruno Rocha 651 Dec 29, 2022
Karte der Allgemeinverfügungen zu Schulschließungen oder eingeschränktem Regelbetrieb in Sachsen

SNSZ Karte Datenquelle: Allgemeinverfügungen zu Schulschließungen oder eingeschränktem Regelbetrieb in Sachsen Sächsisches Staatsministerium für Kultu

Jannis Leidel 3 Sep 26, 2022
This program tries to book a tennis court slot in either Southwark Park or Tanner Street Park in Southwark, London.

Book tennis courts in London This program tries to book a tennis court slot in either Southwark Park or Tanner Street Park in Southwark, London. Note:

Daniele 1 Jul 25, 2022
Synthetik Python Mod - A save editor tool for the game Synthetik written in python

Synthetik_Python_Mod A save editor tool for the game Synthetik written in python

2 Sep 10, 2022
Blender-3D-SH-Dma-plugin - Import and export Sonic Heroes Delta Morph animations (.anm) into Blender 3D

io_scene_sonic_heroes_dma This plugin for Blender 3D allows you to import and ex

Psycrow 3 Mar 22, 2022
Hands-on machine learning workshop

emb-ntua-workshop This workshop discusses introductory concepts of machine learning and data mining following a hands-on approach using popular tools

ISSEL Soft Eng Team 12 Oct 30, 2022
ASVspoof 2021 Baseline Systems

ASVspoof 2021 Baseline Systems Baseline systems are grouped by task: Speech Deepfake (DF) Logical Access (LA) Physical Access (PA) Please find more de

91 Dec 28, 2022
A collection of repositories used to realise various end-to-end high-level synthesis (HLS) flows centering around the CIRCT project.

circt-hls What is this?: A collection of repositories used to realise various end-to-end high-level synthesis (HLS) flows centering around the CIRCT p

29 Dec 14, 2022
A free website that keeps the people informed about housing and evictions.

Eviction Tracker Currently helping verify detainer warrant data for middle Tennessee - via Middle TN DSA - Red Door Collective Features Presents data

Red Door Collective 7 Dec 14, 2022
Collection of functions for working with interlaced content in VapourSynth.

vsfieldkit Collection of functions for working with interlaced content in VapourSynth. It does not have any hard dependencies outside of VapourSynth.

Justin Turner Arthur 11 May 27, 2022
Processamento da Informação - Disciplina UFABC

Processamento da Informacao Disciplina UFABC, Linguagem de Programação Python - 2021.2 Objetivos Apresentar os fundamentos sobre manipulação e tratame

Melissa Junqueira de Barros Lins 1 Jun 12, 2022
Python-geoarrow - Storing geometry data in Apache Arrow format

geoarrow Storing geometry data in Apache Arrow format Installation $ pip install

Joris Van den Bossche 11 Mar 03, 2022
Um pequeno painel de consulta grátis.

[PAINEL-DE-CONSULTA 3.8(BETA)] · Confira meu canal do YouTube. Clique aqui! Nota: Próxima Atualização será a última com coisas novas, o resto será par

276 Jan 05, 2023
Buggy script to play with GPOs

GPOwned /!\ This is a buggy PoC I made just to play with GPOs in my lab. Don't use it in production! /!\ The script uses impacket and ldap3 to update

45 Dec 15, 2022
Certipy is a Python tool to enumerate and abuse misconfigurations in Active Directory Certificate Services (AD CS).

Certipy Certipy is a Python tool to enumerate and abuse misconfigurations in Active Directory Certificate Services (AD CS). Based on the C# variant Ce

ollypwn 1.3k Jan 01, 2023
Drug Discovery App Using Lipinski's Rule-of-Five.

Drug Discovery App A Drug Discovery App Using Lipinski's Rule-of-Five. TAPIWA CHAMBOKO 🚀 About Me I'm a full stack developer experienced in deploying

tapiwa chamboko 3 Nov 08, 2022
Movie recommend community

README 0. 초록 1) 목적 사용자의 Needs를 기반으로 영화를 추천해주는 커뮤니티 서비스 구현 2) p!ck 서비스란? "pick your taste!" 취향대로 영화 플레이리스트(이하 서비스 내에서의 명칭인 '바스켓'이라 함)를 만들고, 비슷한 취향을 가진

2 Dec 08, 2021
PDX Code Guild Full Stack Python Bootcamp starting 2022/02/28

Class Liger Rough Timeline Weeks 1, 2, 3, 4: Python Weeks 5, 6, 7, 8: HTML/CSS/Flask Weeks 9, 10, 11: Javascript Weeks 12, 13, 14, 15: Django Weeks 16

PDX Code Guild 5 Jul 05, 2022