AI Text-to-Video Tutorial: zeroscope_v2_XL in Automatic1111 + tips and tricks DigThatData 2023-06-25 | NB: This tutorial assumes you have automatic1111's "stable diffusion webui" installed already. On my setup, webui is installed inside of docker, but this is NOT a requirement for this tutorial. You just need to have webui installed *somewhere*. Installing docker and configuring it to talk to your GPU can be challenging.vanilla webui: github.com/AUTOMATIC1111/stable-diffusion-webui########################################### Links Mentioned in Video ######################zeroscope_v2_XL announcement tweet: twitter.com/cerspense/status/1672365482454958080Original modelscope files: huggingface.co/damo-vilab/modelscope-damo-text-to-video-synthesis/tree/mainBase model: huggingface.co/cerspense/zeroscope_v2_576w/tree/mainUpscaler: huggingface.co/cerspense/zeroscope_v2_XL/tree/mainThread where I shared about repeatedly applying the upscaler: twitter.com/DigThatData/status/1672542830437470208##########################################Me: twitter.com/DigThatData#ai #aiart #aiartcommunity #texttovideoai #text2image #stablediffusion
Testing VKTRS MSA segmentation for theme assignment DigThatData 2023-05-24 | colab.research.google.com/github/dmarx/video-killed-the-radio-star/blob/main/Video_Killed_The_Radio_Star_Defusion.ipynb
The Living Moon (SD+KLMC2) DigThatData 2023-03-21 | moon 351678610 768 24fps g0 05 sigma2 0 h0 05 gamma1 4 alpha0 001 tau1 0 steps7 hvp fake n1000
Superposition Pizza (KLMC2) DigThatData 2023-03-11 | Couldn't decide what kind of pizza to get, so I ordered a superposition of every pizza on the menu.twitter.com/DigThatData/status/1632301849847169024
PSA: avoid looking in the mirror while on psychedelics DigThatData 2023-02-18 | KLMC2 dreaming on my profile picture, looping between two prompts generated by passing my pfp through PEZ ("hard prompts made easy")
KLMC2 watercolor fantasy DigThatData 2023-02-15 | prompt: "watercolor painted by Coldplay, best of show, photorealistic fantasy, deepdream, excellent use of color"
Organic Architecture 2 - KLMC2 dreaming, 7mins DigThatData 2023-02-07 | 2023 02 07 organic architecture 5kframes 12fps upscaled (for youtube compression)
Organic Architecture: #stablediffusion KLMC2 dreaming about the Sagrada Familia DigThatData 2023-02-07 | #stablediffusion KLMC2 dreaming a transition between two permutations of a Sagrada Familia prompt, looping through 4 ecological prompts at different frequencies, MCMC step and friction params"about" = "in the vicinity of"Get it? "Dreaming About..."? Aren't I super clever?
KLMC2 dreaming eye DigThatData 2023-01-29 | Init image made with DeepFloyd IFVideo made with colab.research.google.com/github/dmarx/notebooks/blob/main/Stable_Diffusion_KLMC2_Animation.ipynb 512x512 2k frames 60fps, upsampled to 1024x1024n = 2000height = 512width = 512seed = 2663381410 init_image = "if - eye 8.jpeg"g = "0:0.085"sigma = "0.5"h = "0:0.07"gamma = "1.3"alpha = "0.005"tau = "1.0"refinement_steps = "6"img_init_steps = 15hvp_method = 'fake'prompt_params = [ ["close up of a woman's eye, resnet convnet filter activations, deepdream, painted by chuck close, pointillism", {0:1}], ["watermark text", {0:-0.1} ], ["jpeg artifacts", {0:-0.1} ],]
Tutorial for VKTRS notebook: make music videos with AI! DigThatData 2022-10-11 | Feature overview and usage tips for my "Video Killed The Radio Star ...Diffusion" notebook.Notebook:colab.research.google.com/github/dmarx/video-killed-the-radio-star/blob/main/Video_Killed_The_Radio_Star_Defusion.ipynbIf you encounter problems, best way to alert me is to create an issue on Github: github.com/dmarx/video-killed-the-radio-starSecond best is to hit me up on twitter: twitter.com/DigThatDataBig thanks to stability.ai for supporting my work, and thanks also to twitter.com/_shadowanderer who edited this down for me from 45min of footage because I talk too damn much.Longer, more detailed edit of this tutorial (28min): youtu.be/xqRAKrmzdCY
