in prep for tomorrow’s first Advanced StyleGAN class I added a bilateral symmetry alignment tool to my dataset-tools library.

0 14

In addition to single-locomotion mode trials, this dataset also includes transitions - including walking to/from ramp ascent, ramp descent, stair ascent, and stair descent.

4 4

TITLE: Latent Aesop 001 DESCRIPTION: Generated with a dataset of etching and engraving illustrations of Aesop's Fables. Artist: https://t.co/zVookq2pTl

2 9

TITLE: haze aesthetics DESCRIPTION: A new aesthetics exploration series. 77 NFTs created using StyleGAN trained on a vaporwave dataset 1024 x 1024. Artist: NEURAL HAZE https://t.co/mALGh907q1

1 2

Large variations in final outcome when 40+ different labs analyzed the same MRI dataset for tractography. Hitting limits of complex technologies. Wouldn't be surprised if similar variability occurred with blind analysis of scRNA-seq datasets. Solutions?

https://t.co/8bp4pVfEMO

20 56

Into the Ether: A Cosmology of the Imaginarium. I think thru things by making images & so new works, a collab between me & the platform where I trained it on a dataset of images of ethereum & then began to layer & collage. Constellations & planets” emerged 🌌

2 15

thanks to and their hard working building a dataset, i was able to train pix2pix to add Garfield back into Garfield Minus Garfield

or at least, attempt to

88 307

4-5 hours of transfer learning from my previous best GAN on a newer dataset

0 0

The dataset from the XES 02 'Jurassic Tank' experiment (that was run a while ago) is fantastic for visualizing how stratigraphy is created during cyclic sea-level change

218 875

https://t.co/pIrbqZGpZc

Here's a peice with my original music. Inspired by any NFT in this style that doesn't credit is pulling a fast one, those folks put in the dataset and training work.

1 2

"Substrate", chemical equilibrium. These visualizations were trained on a dataset of graphical line elements and 3D shape images. Using generative adversarial networks (GAN) https://t.co/N9gDFlRdoi

0 0

"Substrate", chemical equilibrium. These visualizations were trained on a dataset of graphical line elements and 3D shape images. Using generative adversarial networks (GAN) https://t.co/N9gDFlRdoi

0 0

Once I had a good correspondence between new and old, I reversed the transformation, taking a modern elevation dataset and warping it into the space of the old map.

0 6

🏝 Galápagos Islands Concept - Been binging on the new SRTM15+ Global Bathy/Topo Datasets. Game changer.
⠀⠀⠀⠀⠀⠀⠀⠀⠀
https://t.co/h0cg1rMU2f
⠀⠀⠀⠀⠀⠀⠀⠀⠀

7 32

https://t.co/tjQPx6fNlk ヤバイね
手足がおかしかったり、髪の毛が身体や背景と融合してたり、まだ奇形が多いけど、細かいとこ無視れば下手な人間絵師より上手い笑

何がおかしいかconsistencyを学習するにはどうすればいいのか…。造形能力は可能性∞

Danbooru2019 Datasetを拡張したStyleGAN2で

0 2

Anyone been out exploring the available in OpenTopography? One of our team just noticed that Devils Tower National Monument is covered in a recently released dataset (https://t.co/jGnFKVCn8T). Images produced w/ OT browser-based tools.

36 157

me and are building some cool opencv tools to help create datasets. I never thought I’d be able to separate all these flowers without doing it by hand. Now its more sorting that lassoing so its prob like 2x efficient.

1 35

Read the 1st paper in on how small protein omics datasets can find novel genes used in 𝘚𝘢𝘭𝘮𝘰𝘯𝘦𝘭𝘭𝘢 infection

https://t.co/xVPsn5s5a5

18 24

working pretty well! (some issues with the photography in many sets but this will still probably 2-4x my dataset efficiency)

0 1

If you fancy some weekend shopping, how about a Rose for your loved one?

AI generated video using a custom watercolour dataset, some beautiful generated piano music and a small poem to go by...

Available on :
https://t.co/TJUcDJweMV

6 14