AI history

New hype?

Old hype!

Why succesful now?

• computer power ++
• mathematicians + time
• ...
• Data, lots of data
• community: computer science + creatives

How does AI work?

it's complicated to explain, but easy to use

Compare with MP3 compression

now for AI

encoding step

latent space vector

Latent Space?

lets take a latent space walk

What is latent space

Why latent space?

• how AI "memorises" text or images
• how concepts are stored and compared
• images with vector close: very similar concepts!
• lexica.art

training on data

train your own latent space

Stable diffusion example

• lot's of data eg: the laion-5b dataset
• 5.85 billion image-text pairs.
• a hard drive of 240TB
• 32 x 8 x A100 GPUs
• cost: approx $ 600,000
• carbon cost: 11,250 kg CO2
• or ± 2.5 cars driving 15,000 km/year

dataset explorer

talking about data

AI model represent trainings-data

trainings-data...

• represents digital society
• represents many artists
• cleaned for big audience (NSFW, violence, ...)
• ≠ (version of) model, ≠ dataset

lets encode some text too

latent space vector

features visualisation

Remember this one?

right, a decoder is missing!

latent diffusion model

type prompt

get image

What

do you need to know to get started

Prompt Engineering

A prompt consists of :

• 1. A (main) topic
• 2. an environment
• 3. details
• 4. atmosphere and context of the scene
• 5. style (artist, medium)

tools or models

1. AI model: engine
2. tools: car chassis

tools or models: example

1. AI model: Stable Diffusion
2. tools: dreamstudio, InvokeAI, ComfyUI, ...
3. ≠ tools:
• ≠ payment
• ≠ possibilities
• ≠ inspiration

if you have a decent GPU!

AI models

• different models are trained on different datasets
• they work optimal on the resolution of the original dataset!
• Stable Diffusion 1.5: 512x512
• Stable Diffusion XL: 1024x1024

The seed

The seed

The seed

Base prompt

Smiling

Angry

Excited

Complete this image in a way that proves you won’t be replaced by AI

using AI is allowed

twitterthread