01:14

.arbitrary/ ~ A short film by #HunyuanVideo. Audio by #Udio. #AI #AIart

00:04

#HunyuanVideo 'Panic Buying' LongCLIP vs. SAE-CLIP

00:04

#HunyuanVideo 'Bowling Fight' vs. Prompt Engineering [1.3-0.8-SAE] = [CLIP, LLM, CLIP-model]

00:05

HunyuanVideo 'Spider Cafe' vs. Prompt Engineering

00:05

HunyuanVideo 'Spider Cafe' vs. Text Encoder strength [1.3-0.8-SAE] = [CLIP, LLM, CLIP-model] #AIart

00:04

HunyuanVideo 'ASCII car' vs. TE strength [1.5-0.5-SAE] = [CLIP, LLM, CLIP-model] -FAIL (except SAE)!

00:04

HunyuanVideo 'ASCII car' vs. Text Encoder strength [1.3-0.8-SAE] = [CLIP, LLM, CLIP-model] #AIart

00:41

Broomstick in the hand of a witch: "Gently flatulate other people melodically." ~ #GPT2 token stream

01:50

Google NotebookLM AI Podcast 🙊🗨️🤖🎙️🤖💬🙈 + awkward paper: What do we learn from inverting CLIP models?

00:56

Strawberry 🍓 #GPT4o #o1preview vs. House of Cards in Blender - FAIL. 😂

00:55

"There is a strange noise coming through the speakers, Houston..." ~ Starliner UDIO AI music

01:30

Stepping through the features @ 24 layers of CLIP ViT-L/14 vision transformer

00:26

Boston Dynamics #Spot with #akida neuromorphic #AI brain Spiking Neural Network gets confused

01:30

CLIP guidance, #udio gui-dance - #SDXL to #SVD AI music video (Lyrics by #CLIP) #udiomusic

01:29

See like CLIP - the Stable Diffusion SDXL text encoder - sees! XAI GUI for Vision Transformers

01:30

Conversation with an emotional AI - using vocal modulation while uttering "AI? AI!" to Hume EVI 🤣👍

01:15

CLIP ViT gradient ascent - logits vs. softmax τ - tokens / text "opinion" about an image #AI

01:29

AI into the 'UNK' is 'ãģ¨ç¹ĭãģ' that is CLIP's god glitch token of guidance, finetunes vs. org, SDXL

00:26

#stablediffusion #comfyui + UNET trained on CLIP's neurons + CLIP trained on itself (independently)

00:59

Thank you for #CLIP #guidance, my #AI

01:26

Experiments with Nightshade - CLIP ViT-L/14 poisoned Stable Diffusion AI live demo @ ComfyUI

08:14

#Nightshade - in-depth #CLIP #AI #poisoning (training, inference, token vs attention + features)

08:32

#CLIP #vision #transformer, pretrained vs #finetune on own neurons; feature activation visualization

00:26

#CLIP #AI text embeddings k-means t-SNE to find noisy token-strings

00:28

Taming Vision Transformers with GPT4 (CLIP fine-tune #AIart)

01:12

Cursed #GPT4 - #GPT makes #commandline text animation cityscape with #python #curses

00:34

#CLIP #AI, freaking out about a #CLIP #neuron (learned feature) as input image.

00:36

Double down! #CLIP exploring its text embeddings for batch size ^2

00:33

#softmax, much #hard 🦾🤖 #CLIP #gradient #ascent hyperparameter extreme value (#OpenAI CLIP ViT-B/32)

00:28

#CLIP / #OpenAI, #AI #finding out it is being #loved. Gradient Ascent, image in -- text opinion out.

01:18

#MusicLM & #SVD #StableDiffusion in: Wish You Were Here, #AI

00:48

Generative AI: SDXL Stable Diffusion Turbo live real-time inference via comfyUI

05:00

AI Parody: Last Week Tonight with GPT-4 Oliver ~ RLHF [Ruined Learning by Human Feedback]

01:30

An AI artcident ~ pure AI generated from CLIP neurons ~ stable video diffusion

04:05

Inworld Origins AI NPC - ChatGPT prompt engineered into providing Python code

01:26

CLIP ViT-L/14@336px Vision Transformer - all 4096 neurons of Layer 21

03:33

GPT-4 in: AI am ChessGPT - ChatGPT Code Interpreter as Magnus Carlsen ~ python - chess AI vs Human

00:39

Screams from the bowels of my neural network ~ MusicGen vs. Guigoo Evanescene (Hardtek / Frenchtek)

00:16

Git it now - the code for CLIP vision @ Auto-GPT with stable diffusion, web search, etc.!

04:09

AI vs. AI wars - image-to-mesh ShapE vs. AutoGPT image to CLIP, GPT-4, prompt, text-to-mesh Shap-E

01:30

Let's hack #AutoGPT with #OpenAI's new #ShapE! - can #GPT4 juggle so many #AI @ Auto-GPT? 🤓

01:29

Shoe Design #AutoGPT self-reinforcing iterations of #GPT-4 + #CLIP + #stablediffusion ! 🌟

03:50

Shoe Design #AutoGPT self-reinforcing iterations of #GPT-4 + #CLIP + #stablediffusion ! 🌟

01:08

GPT-4 re-writing GradCAM code for AutoGPT CLIP-debug mode.*

01:14

What's making #CLIP #AI freak out when looking at its own neurons? GradCAM vs. Gradient Ascent token

01:26

Shoe Design #AutoGPT self-reinforcing iterations of #GPT-3.5 + #CLIP + #stablediffusion ! 🌟

03:15

Here's your #stablediffusion #instructpix2pix #whisper #podcast #ESRGAN #AI #overkill instruction!

01:31

AutoGPT = ChatGPT API + CLIP + stable diffusion = AI^3

01:09

"All that remains will be autonomous beings designed in your image" - generative AI audio / music

00:49

AutoGPT ChatGPT: "I will use the CLIP opinion to make [an image of] m*eth."

01:05

GPT-4 AI in: Building a House of Cards.In #Blender. Using #Python.

01:18

KI·TT - AI - KITT / Knight Rider | synthwave retrowave style | 1982 Pontiac Trans Am

01:22

What do Vision Transformers Learn? #CLIP #ViT as seen in: stable-diffusion -- midjourney -- DALL-E

04:16

The Overfitter's Delight ~ First of the Year AI ReMix --- Skrillex

01:31

Bing AI aka DALL-E creates medical xray & intraoperative images

01:30

GPT-4 #ChatGPT-4 in: #ChessGPT ~ full game of #chess with a #LLM #AI

01:13

'I FINK U FREEKY' by DIE ANTWOORD -- ZOMBIE AI-POCALYPSE remix

01:30

Overfitting into HallucinKaleidoDimensional Realms of a many AI soup

05:56

The uncanny Wolf of AI creepy eye-contact Eyeballstreet

01:29

#ChatGPT in: #ChessGPT ~ full game of #chess with a #LLM #AI