more translation
This commit is contained in:
parent
2ce9ca50b5
commit
a41be821c1
997 changed files with 33247 additions and 32490 deletions
|
|
@ -1,20 +1,20 @@
|
|||
+++
|
||||
title = "Übersetzung: Local Diffusion"
|
||||
excerpt = "Empower your own Stable Diffusion Generation: InKüLe supported student workshop: Local Diffusion by Aron Petau"
|
||||
title = "Lokale Diffusion"
|
||||
excerpt = "Empower your own Stable Diffusion Generation: InKüLe unterstützter studentischer Workshop von Aron Petau"
|
||||
date = 2024-04-11
|
||||
authors = ["Aron Petau"]
|
||||
banner = "images/local-diffusion/local-diffusion.png"
|
||||
banner = "images/local-diffusion/local-diffusion.png"
|
||||
description = "Ein praxisorientierter Workshop zur ethischen KI-Bilderzeugung durch lokale Stable Diffusion, bei dem Studierende lernen, Graphic Novels zu erstellen und die Bedeutung von lokalem Computing und Datensouveränität zu verstehen."
|
||||
|
||||
[taxonomies]
|
||||
tags = [
|
||||
"automatic1111",
|
||||
"stable diffusion",
|
||||
"comfyui",
|
||||
"diffusionbee",
|
||||
"inküle",
|
||||
"Local Computing",
|
||||
"Stable Diffusion",
|
||||
"Workshop",
|
||||
"university of the arts berlin"
|
||||
"workshop",
|
||||
"university of the arts berlin",
|
||||
]
|
||||
|
||||
[extra]
|
||||
|
|
@ -22,23 +22,147 @@ show_copyright = true
|
|||
show_shares = true
|
||||
+++
|
||||
|
||||
## Local Diffusion
|
||||
## Kernfragen
|
||||
|
||||
[The official call for the Workshop](https://www.udk-berlin.de/universitaet/online-lehre-an-der-universitaet-der-kuenste-berlin/inkuele/11-april-24-aron-stable-diffusion/)
|
||||
Ist es möglich, eine Graphic Novel mit generativer KI zu erstellen?
|
||||
Was bedeutet es, diese neuen Medien in Zusammenarbeit mit anderen zu nutzen?
|
||||
Und warum sind ihre lokalen und offline-Anwendungen wichtig?
|
||||
|
||||
Is it possible to create a graphic novel with generative A.I.?
|
||||
What does it mean to use these emerging media in collaboration with others?
|
||||
And why does their local and offline application matter?
|
||||
[Offizielle Workshop-Dokumentation](https://www.inkuele.de/dokumentation/details/lokale-diffusion-br-empower-your-own-em-stable-diffusion-generation-em) | [Workshop-Ausschreibung](https://www.udk-berlin.de/universitaet/online-lehre-an-der-universitaet-der-kuenste-berlin/inkuele/11-april-24-aron-stable-diffusion/)
|
||||
|
||||
With AI becoming more and more democratised and GPT-like Structures increasingly integrated into everyday life, the black-box notion of the mysterious all-powerful Intelligence hinders insightful and effective usage of emerging tools. One particularly hands-on example is AI generated images. Within the proposed Workshop, we will dive into Explainable AI, explore Stable Diffusion, and most importantly, understand the most important parameters within it. We want to steer outcomes in a deliberate manner. Emphasis here is on open and accessible technology, to increase user agency and make techno-social dependencies and power relations visible.
|
||||
## Workshop-Ziele & Struktur
|
||||
|
||||
Empower yourself against readymade technology!
|
||||
Do not let others decide on what your best practices are. Get involved in the modification of the algorithm and get surprised by endless creative possibilities. Through creating a short graphic novel with 4-8 panels, participants will be able to utilise multiple flavours of the Stable Diffusion algorithm, and will have a non-mathematical understanding of the parameters and their effects on the output within some common GUIs. They will be able to apply several post-processing techniques to their generated images, such as upscaling, masking, inpainting and pose redrawing. Further, participants will be able to understand the structure of a good text prompt, be able to utilise online reference databases and manipulate parameters and directives of the Image to optimise desired qualities. Participants will also be introduced to ControlNet, enabling them to direct Pose and Image composition in detail.
|
||||
### Fokus: Theoretische und spielerische Einführung in A.I.-Tools
|
||||
|
||||
## Workshop Evaluation
|
||||
Der Workshop verfolgte ein doppeltes Ziel:
|
||||
|
||||
Over the course of 3 hours, I gave an introductory workshop in local stable diffusion processing and introduced participants to the server available to UdK Students for fast remote computation that circumvents the unethicality of continuously using a proprietary cloud service for similar outputs. There is not much we can do on the data production side and many ethical dilemmas surrounding digital colonialism remain, but local computation takes one step towards a critical and transparent use of AI tools by Artists.
|
||||
1. **Niedrigschwelliger Einstieg**: Anfänger*innen einen zugänglichen Einstieg in die Text-to-Image-KI ermöglichen
|
||||
2. **Kritische Diskussion**: Eine differenzierte politische Diskussion über die ethischen Auswirkungen dieser Tools anstoßen und bewusste Entscheidungsoptionen aufzeigen (wie lokal installierte Tools)
|
||||
|
||||
The Workshop format was rathert open and experimental, which was welcomed by the participants and they tried the collages enthusiastically. We also had a refreshing discussion on different positions regarding the ethicalities and whether a complete block of these tools is called for and feasible.
|
||||
Das Lernformat wurde offen, praxisnah und experimentell gestaltet, wobei der Schwerpunkt auf dem kreativen Output der Teilnehmer*innen lag. Konkret wurden sie aufgefordert, in Gruppen zu arbeiten und gemeinsam mit der KI eine kurze Graphic Novel mit 4-8 Panels zu erstellen. Dabei mussten sie den Algorithmus aktiv verändern und sich mit den verschiedenen Funktionen und Schnittstellen vertraut machen.
|
||||
|
||||
I am looking forward to round 2 with the next iteration, where we are definitely diving deeper into the depths of comfyui, an interface that i absolutely adore, while its power also terrifies me sometimes.
|
||||
### Workshop-Ablauf
|
||||
|
||||
Der Workshop war in zwei Hauptteile gegliedert:
|
||||
|
||||
#### Teil 1: Theoretische Einführung (45 Min.)
|
||||
|
||||
- Entmystifizierung der Prozesse, die im Hintergrund ablaufen
|
||||
- Einführung in den Stable Diffusion Algorithmus
|
||||
- Verständnis des Diffusionsprozesses und der Noise Reduction
|
||||
- Unterschiede zu älteren Generative Adversarial Networks (GANs)
|
||||
- Ethische Implikationen des Einsatzes von KI-Tools
|
||||
|
||||
#### Teil 2: Praktische Übungen (2+ Stunden)
|
||||
|
||||
- "Stadt-Land-Fluss"-Spiel zur Prompt-Konstruktion
|
||||
- Erstellung einer Graphic Novel mit 4-8 Panels
|
||||
- Experimentieren mit Parametern und Schnittstellen
|
||||
- Nachbearbeitungstechniken (Upscaling, Maskieren, Inpainting, Pose Redrawing)
|
||||
- Gruppenpräsentationen und Diskussion
|
||||
|
||||
### Das "Stadt-Land-Fluss"-Aufwärmspiel
|
||||
|
||||
Um die anfängliche Angst vor dem Prompting zu überwinden, spielten die Teilnehmer*innen eine Runde "Stadt-Land-Fluss" (Kategorien). Sie mussten vordefinierte Prompting-Unterkategorien wie "Thema", "Farbe", "Stil" und "Auflösung" mit Worten füllen, die mit bestimmten Buchstaben beginnen. Dieses Spiel fordert die Teilnehmenden heraus, sich in die kreative Gestaltung eines Prompts hineinzudenken, jenseits von vorgefertigten Sätzen, wie sie online zu finden sind.
|
||||
|
||||
## Warum lokale KI-Tools verwenden?
|
||||
|
||||
### Bewusst ethische und datenschutzrechtliche Faktoren miteinbeziehen
|
||||
|
||||
Eine zentrale Idee des Workshops war, die ethischen Implikationen des Einsatzes von KI-Tools in den Fokus zu rücken und Konsequenzen von lokaler Rechenleistung im Gegensatz zum Cloud-Computing hervorzuheben. Der Workshop thematisierte zwei wesentliche Unterschiede bei der Anwendung derselben KI-Modelle und -Algorithmen:
|
||||
|
||||
#### Option 1: Proprietäre Cloud-Dienste
|
||||
|
||||
- Populäre Plattformen wie Midjourney
|
||||
- Schnittstelle von privaten Unternehmen bereitgestellt
|
||||
- Oft gebührenpflichtig
|
||||
- Ergebnisse auf Unternehmensservern gespeichert
|
||||
- Daten für weiteres KI-Modell-Training verwendet
|
||||
- Begrenzte Benutzerkontrolle und Transparenz
|
||||
|
||||
#### Option 2: Lokale Installation
|
||||
|
||||
- Selbst installierte Apps auf privaten Computern
|
||||
- Selbst installierte GUIs oder Front-Ends über Browser zugänglich
|
||||
- Vollständige Datensouveränität
|
||||
- Keine Datenweitergabe an Dritte
|
||||
- Offline-Fähigkeit
|
||||
|
||||
#### Option 3: Universitäts-gehostete Dienste
|
||||
|
||||
- Transparente Anbieter (z.B. UdK Berlin Server)
|
||||
- Schneller und zuverlässiger als proprietäre Cloud-Dienste
|
||||
- Daten weder an Dritte weitergegeben noch für Training verwendet
|
||||
- Besser als proprietäre Dienste bei gleichzeitiger Zugänglichkeit
|
||||
|
||||
**Aus Perspektive des Datenschutzes sind lokale und universitäts-gehostete Lösungen bei weitem die bewussteren Wahlen.** Auch wenn UdK-Dienste technisch gesehen ebenfalls Cloud-Dienste mit auf einem Server gespeicherten Daten sind, stellen sie einen großen Unterschied zur Nutzung proprietärer Dienste wie OpenAI dar.
|
||||
|
||||
## Visuelles Erzählen mit Stable Diffusion
|
||||
|
||||
Die Teilnehmer*innen haben sich mit großer Begeisterung auf den Workshop-Prozess eingelassen. Sie probierten viele verschiedene Prompts und Einstellungen aus und produzierten Ergebnisse mit einer großen Vielfalt an ästhetischen und visuellen Erzählungen.
|
||||
|
||||
Der Workshop endete mit einer abschließenden Diskussion über:
|
||||
|
||||
- Die ethischen Implikationen des Einsatzes von KI-Tools
|
||||
- Die Auswirkungen auf die verschiedenen kreativen Disziplinen
|
||||
- Die Frage, ob eine vollständige Abschaffung dieser Tools notwendig oder überhaupt machbar ist
|
||||
|
||||
## Technischer Rahmen
|
||||
|
||||
Mit zunehmender Demokratisierung von KI und der Integration GPT-ähnlicher Strukturen in den Alltag behindert die Black-Box-Vorstellung der mysteriösen allmächtigen Intelligenz die aufschlussreiche und effektive Nutzung aufkommender Tools. Ein besonders praxisnahes Beispiel sind KI-generierte Bilder.
|
||||
|
||||
### Vorgestellte Tools & Schnittstellen
|
||||
|
||||
- **Stable Diffusion**: Der Kern-Algorithmus
|
||||
- **ComfyUI**: Node-basiertes Front-End für Stable Diffusion
|
||||
- **automatic1111**: GUI verfügbar auf UdK Berlin Servern
|
||||
- **DiffusionBee**: Lokale Anwendungsoption
|
||||
- **ControlNet**: Für detaillierte Pose- und Kompositionskontrolle
|
||||
|
||||
### Lernergebnisse
|
||||
|
||||
Die Teilnehmer*innen erlangten die Fähigkeit:
|
||||
|
||||
- Mehrere Varianten des Stable Diffusion Algorithmus zu nutzen
|
||||
- Ein nicht-mathematisches Verständnis von Parametern und deren Effekten zu entwickeln
|
||||
- Nachbearbeitungstechniken anzuwenden (Upscaling, Maskieren, Inpainting, Pose Redrawing)
|
||||
- Effektive Text-Prompts zu konstruieren
|
||||
- Online-Referenzdatenbanken zu nutzen
|
||||
- Parameter zu manipulieren, um gewünschte Qualitäten zu optimieren
|
||||
- ControlNet für detaillierte Pose- und Kompositionssteuerung zu verwenden
|
||||
|
||||
## Erfahrungsbericht von Aron Petau
|
||||
|
||||
### Die Student-als-Lehrer Perspektive
|
||||
|
||||
#### Über Vorbereitung und Herausforderungen
|
||||
|
||||
"Die Vorbereitung eines Workshops fühlte sich definitiv wie eine große Aufgabe an, weil ich das Bedürfnis hatte, Fragen zu Tools zu beantworten, die ich selbst gerade erst entdecke. Eine Sorge war, dass ich die Antwort auf ein fortgeschrittenes technisches Problem nicht geben kann. Dies stellte sich letztendlich als kein großes Problem heraus, was wahrscheinlich an der begrenzten Dauer des Workshops lag.
|
||||
|
||||
Was die Erfahrung mit einem KI-Workshop angeht, so bin ich der Meinung, dass es mehr als 3 Stunden braucht, um gemeinsam mit den Menschen in solche komplexen Werkzeuge einzutauchen. Selbst durch die Ausweitung des erklärenden/theoretischen Teils habe ich es nicht geschafft, alle Konzepte abzudecken, die ich im Vorfeld für wertvoll eingestuft habe... Dennoch erscheint mir die Dauer von 3–4 Stunden für einen Einführungsworkshop angemessen, da sich bei längeren Zeitspannen Fehler im Zeitmanagement summieren und hier vielleicht auch mehr Lehrerfahrung nötig wäre."
|
||||
|
||||
#### Über Workshop-Format und Atmosphäre
|
||||
|
||||
"Gut gefallen hat mir der eher hierarchiearme Rahmen des Workshops, bei dem klar war, dass es sich eher um ein Skillsharing und nicht um ein Vorlesungsformat handelt. Vor allem bei so praktischen Dingen wie der Bilderzeugung konnte ich, wenn ich die Wirkung eines Promptes oder von einem Parameter nicht kannte – wie auch, das ist ja Sinn der Sache – den Effekt einfach gemeinsam mit den Workshop-Teilnehmer*innen ausprobieren und dann die Ergebnisse untersuchen.
|
||||
|
||||
Die Teilnehmer*innen schienen das gewählte Format und den Schwierigkeitsgrad zu mögen, bei dem nicht zu viel Mathematik und Formeln vermittelt wurden, sondern eine Intuition für den zugrunde liegenden Prozess. Die Teilnehmer*innen beteiligten sich auch aktiv an der kritischen Diskussion über den ethischen Einsatz von KI und brachten Perspektiven aus ihren eigenen Bereichen ein, was ich sehr zu schätzen wusste."
|
||||
|
||||
#### Über das Erlernen didaktischer Praxis
|
||||
|
||||
"Während der Vorbereitung dieses Workshops hatte ich die Möglichkeit, selbständig zu arbeiten und meine Workshop-Termine selbst zu bestimmen und zu organisieren. Diese Freiheit und Autorität habe ich sehr geschätzt, aber ein etwas stärkerer Druck auf einen endgültigen Termin hätte mir geholfen, die Bedenken bezüglich der Lehrsituation schneller zu verlieren.
|
||||
|
||||
Jetzt freue ich mich auf eine mögliche Runde 2 – eine nächste Iteration, in der wir tiefer in die Tiefen von ComfyUI eintauchen können, einer Schnittstelle, die ich absolut liebe, während ihre Macht mir manchmal auch Angst macht."
|
||||
|
||||
## Empowerment durch Verständnis
|
||||
|
||||
**Empower yourself against readymade technology!**
|
||||
|
||||
Lass nicht andere darüber entscheiden, was deine Best Practices sind. Beteilige dich an der Modifikation des Algorithmus und lass dich von endlosen kreativen Möglichkeiten überraschen. Durch die Erkundung lokaler KI-Tools können wir:
|
||||
|
||||
- Schritte hin zu einer kritischen und transparenten Nutzung von KI-Tools durch Künstler*innen gehen
|
||||
- Die Handlungsmacht der Nutzer*innen erhöhen
|
||||
- Techno-soziale Abhängigkeiten und Machtverhältnisse sichtbar machen
|
||||
- Fragen des digitalen Kolonialismus ansprechen
|
||||
- Datensouveränität und Privatsphäre bewahren
|
||||
|
||||
Während wir auf der Datenproduktionsseite nicht viel tun können und viele ethische Dilemmata rund um den digitalen Kolonialismus bestehen bleiben, ist lokales Computing ein Schritt hin zu einer kritischen und transparenten Nutzung von KI-Tools durch Künstler*innen.
|
||||
|
|
|
|||
|
|
@ -1,20 +1,19 @@
|
|||
+++
|
||||
title = "Local Diffusion"
|
||||
excerpt = "Empower your own Stable Diffusion Generation: InKüLe supported student workshop: Local Diffusion by Aron Petau"
|
||||
date = 2024-04-11
|
||||
authors = ["Aron Petau"]
|
||||
banner = "images/local-diffusion/local-diffusion.png"
|
||||
description = "A hands-on workshop exploring ethical AI image generation through local Stable Diffusion, teaching students to create graphic novels while understanding the importance of local computing and data sovereignty."
|
||||
|
||||
[taxonomies]
|
||||
tags = [
|
||||
"automatic1111",
|
||||
"stable diffusion",
|
||||
"comfyui",
|
||||
"diffusionbee",
|
||||
"inküle",
|
||||
"Local Computing",
|
||||
"Stable Diffusion",
|
||||
"Workshop",
|
||||
"university of the arts berlin"
|
||||
"workshop",
|
||||
"university of the arts berlin",
|
||||
]
|
||||
|
||||
[extra]
|
||||
|
|
@ -22,23 +21,145 @@ show_copyright = true
|
|||
show_shares = true
|
||||
+++
|
||||
|
||||
## Local Diffusion
|
||||
|
||||
[The official call for the Workshop](https://www.udk-berlin.de/universitaet/online-lehre-an-der-universitaet-der-kuenste-berlin/inkuele/11-april-24-aron-stable-diffusion/)
|
||||
## Core Questions
|
||||
|
||||
Is it possible to create a graphic novel with generative A.I.?
|
||||
What does it mean to use these emerging media in collaboration with others?
|
||||
And why does their local and offline application matter?
|
||||
|
||||
With AI becoming more and more democratised and GPT-like Structures increasingly integrated into everyday life, the black-box notion of the mysterious all-powerful Intelligence hinders insightful and effective usage of emerging tools. One particularly hands-on example is AI generated images. Within the proposed Workshop, we will dive into Explainable AI, explore Stable Diffusion, and most importantly, understand the most important parameters within it. We want to steer outcomes in a deliberate manner. Emphasis here is on open and accessible technology, to increase user agency and make techno-social dependencies and power relations visible.
|
||||
[Official Workshop Documentation](https://www.inkuele.de/dokumentation/details/lokale-diffusion-br-empower-your-own-em-stable-diffusion-generation-em) | [Workshop Call](https://www.udk-berlin.de/universitaet/online-lehre-an-der-universitaet-der-kuenste-berlin/inkuele/11-april-24-aron-stable-diffusion/)
|
||||
|
||||
Empower yourself against readymade technology!
|
||||
Do not let others decide on what your best practices are. Get involved in the modification of the algorithm and get surprised by endless creative possibilities. Through creating a short graphic novel with 4-8 panels, participants will be able to utilise multiple flavours of the Stable Diffusion algorithm, and will have a non-mathematical understanding of the parameters and their effects on the output within some common GUIs. They will be able to apply several post-processing techniques to their generated images, such as upscaling, masking, inpainting and pose redrawing. Further, participants will be able to understand the structure of a good text prompt, be able to utilise online reference databases and manipulate parameters and directives of the Image to optimise desired qualities. Participants will also be introduced to ControlNet, enabling them to direct Pose and Image composition in detail.
|
||||
## Workshop Goals & Structure
|
||||
|
||||
## Workshop Evaluation
|
||||
### Focus: Theoretical and Playful Introduction to A.I. Tools
|
||||
|
||||
Over the course of 3 hours, I gave an introductory workshop in local stable diffusion processing and introduced participants to the server available to UdK Students for fast remote computation that circumvents the unethicality of continuously using a proprietary cloud service for similar outputs. There is not much we can do on the data production side and many ethical dilemmas surrounding digital colonialism remain, but local computation takes one step towards a critical and transparent use of AI tools by Artists.
|
||||
The workshop pursued a dual objective:
|
||||
|
||||
The Workshop format was rathert open and experimental, which was welcomed by the participants and they tried the collages enthusiastically. We also had a refreshing discussion on different positions regarding the ethicalities and whether a complete block of these tools is called for and feasible.
|
||||
1. **Accessible Entry Point**: Provide beginners with a low-barrier introduction to text-to-image AI
|
||||
2. **Critical Discussion**: Initiate a nuanced political discussion about the ethical implications of these tools and demonstrate conscious decision-making options (such as locally installed tools)
|
||||
|
||||
I am looking forward to round 2 with the next iteration, where we are definitely diving deeper into the depths of comfyui, an interface that i absolutely adore, while its power also terrifies me sometimes.
|
||||
The learning format was designed to be open, practical, and experimental, with emphasis on participants' creative output. Specifically, participants were tasked with working in groups to create a short graphic novel of 4-8 panels using AI. They had to actively modify the algorithm and familiarize themselves with various functions and interfaces.
|
||||
|
||||
### Workshop Structure
|
||||
|
||||
The workshop was divided into two main parts:
|
||||
|
||||
#### Part 1: Theoretical Introduction (45 min)
|
||||
|
||||
- Demystifying AI processes running in the background
|
||||
- Introduction to the Stable Diffusion algorithm
|
||||
- Understanding the diffusion process and noise reduction
|
||||
- Differences from older Generative Adversarial Networks (GANs)
|
||||
- Ethical implications of AI tool usage
|
||||
|
||||
#### Part 2: Hands-On Practice (2+ hours)
|
||||
|
||||
- "Categories Game" for prompt construction
|
||||
- Creating a 4-8 panel graphic novel
|
||||
- Experimenting with parameters and interfaces
|
||||
- Post-processing techniques (upscaling, masking, inpainting, pose redrawing)
|
||||
- Group presentations and discussion
|
||||
|
||||
### The "Categories Game" Warm-Up
|
||||
|
||||
To overcome initial fears about prompting, participants played a round of "Categories" (Stadt-Land-Fluss). They had to fill predefined prompting subcategories like "Subject", "Color", "Style", and "Resolution" with words starting with specific letters. This game challenged participants to think creatively about prompt construction beyond ready-made sentences found online.
|
||||
|
||||
## Why Local AI Tools Matter
|
||||
|
||||
### Consciously Considering Ethical and Data Protection Factors
|
||||
|
||||
A central idea of the workshop was to highlight the ethical implications of using AI tools and emphasize the consequences of local computing versus cloud computing. The workshop addressed two essential differences when applying the same AI models and algorithms:
|
||||
|
||||
#### Option 1: Proprietary Cloud Services
|
||||
|
||||
- Popular platforms like Midjourney
|
||||
- Interface provided by private companies
|
||||
- Often fee-based
|
||||
- Results stored on company servers
|
||||
- Data used for further AI model training
|
||||
- Limited user control and transparency
|
||||
|
||||
#### Option 2: Local Installation
|
||||
|
||||
- Self-installed apps on private computers
|
||||
- Self-installed GUIs or front-ends accessed via browser
|
||||
- Complete data sovereignty
|
||||
- No third-party data sharing
|
||||
- Offline capability
|
||||
|
||||
#### Option 3: University-Hosted Services
|
||||
|
||||
- Transparent providers (e.g., UdK Berlin servers)
|
||||
- Faster and more reliable than proprietary cloud services
|
||||
- Data neither shared with third parties nor used for training
|
||||
- Better than proprietary services while maintaining accessibility
|
||||
|
||||
**From a data protection perspective, local and university-hosted solutions are far more conscious choices.** While UdK services are technically also cloud services with data stored on a server, they represent a significant difference from proprietary services like OpenAI.
|
||||
|
||||
## Visual Storytelling with Stable Diffusion
|
||||
|
||||
Participants engaged enthusiastically in the workshop process. They tried many different prompts and settings, producing results with a great variety of aesthetic and visual narratives. The workshop concluded with a final discussion about:
|
||||
|
||||
- Ethical implications of using AI tools
|
||||
- Impact on various creative disciplines
|
||||
- Whether complete abolition of these tools is necessary or even feasible
|
||||
|
||||
## Technical Framework
|
||||
|
||||
With AI becoming increasingly democratized and GPT-like structures integrated into everyday life, the black-box notion of the mysterious all-powerful intelligence hinders insightful and effective usage of emerging tools. One particularly hands-on example is AI-generated images.
|
||||
|
||||
### Tools & Interfaces Introduced
|
||||
|
||||
- **Stable Diffusion**: The core algorithm
|
||||
- **ComfyUI**: Node-based front-end for Stable Diffusion
|
||||
- **automatic1111**: GUI available on UdK Berlin servers
|
||||
- **DiffusionBee**: Local application option
|
||||
- **ControlNet**: For detailed pose and composition control
|
||||
|
||||
### Learning Outcomes
|
||||
|
||||
Participants gained the ability to:
|
||||
|
||||
- Utilize multiple flavors of the Stable Diffusion algorithm
|
||||
- Develop non-mathematical understanding of parameters and their effects
|
||||
- Apply post-processing techniques (upscaling, masking, inpainting, pose redrawing)
|
||||
- Construct effective text prompts
|
||||
- Utilize online reference databases
|
||||
- Manipulate parameters to optimize desired qualities
|
||||
- Use ControlNet for detailed pose and composition direction
|
||||
|
||||
## Reflections: The Student-as-Teacher Perspective
|
||||
|
||||
*Personal reflection by Aron Petau*
|
||||
|
||||
### On Preparation and Challenges
|
||||
|
||||
"Preparing a workshop definitely felt like a big task because I felt the need to answer questions about tools that I myself am just discovering. One concern was that I wouldn't be able to answer an advanced technical problem. This ultimately turned out not to be a major issue, probably due to the limited duration of the workshop.
|
||||
|
||||
When it comes to the experience with an AI workshop, I believe it takes more than 3 hours to dive into such complex tools together with people. Even by extending the explanatory/theoretical part, I didn't manage to cover all the concepts I had deemed valuable beforehand... Nevertheless, a duration of 3-4 hours seems appropriate for an introductory workshop, as errors in time management accumulate over longer periods and more teaching experience would be needed here."
|
||||
|
||||
### On Workshop Format and Atmosphere
|
||||
|
||||
"I really liked the rather non-hierarchical framework of the workshop, where it was clear that it was more about skill-sharing rather than a lecture format. Especially with practical things like image generation, when I didn't know the effect of a prompt or a parameter – which is the point after all – I could simply try out the effect together with the workshop participants and then examine the results.
|
||||
|
||||
The participants seemed to like the chosen format and difficulty level, where not too much mathematics and formulas were conveyed, but rather an intuition for the underlying process. The participants also actively participated in the critical discussion about the ethical use of AI and contributed perspectives from their own fields, which I greatly appreciated."
|
||||
|
||||
### On Learning Teaching Practice
|
||||
|
||||
"During the preparation of this workshop, I had the opportunity to work independently and determine and organize my workshop dates myself. I greatly appreciated this freedom and authority, but a bit stronger pressure on a final deadline would have helped me lose the concerns about the teaching situation more quickly.
|
||||
|
||||
Now I'm looking forward to a possible round 2 – a next iteration where we can dive deeper into the depths of ComfyUI, an interface that I absolutely love, while its power also sometimes scares me."
|
||||
|
||||
## Empowerment Through Understanding
|
||||
|
||||
**Empower yourself against ready-made technology!**
|
||||
|
||||
Do not let others decide on what your best practices are. Get involved in the modification of the algorithm and get surprised by endless creative possibilities. Through exploring local AI tools, we can:
|
||||
|
||||
- Take steps toward critical and transparent use of AI tools by artists
|
||||
- Increase user agency
|
||||
- Make techno-social dependencies and power relations visible
|
||||
- Address issues of digital colonialism
|
||||
- Maintain data sovereignty and privacy
|
||||
|
||||
While there is not much we can do on the data production side and many ethical dilemmas surrounding digital colonialism remain, local computation takes one step towards a critical and transparent use of AI tools by artists.
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue