Hugging Face's logo

ivanmikhnenkov
/
tinydit-256

Text-to-Image
Diffusers
Safetensors
diffusion-transformer
rectified-flow
from-scratch
Model card Files
xet
 Community
1
ivanmikhnenkov commited on
Commit
f330fd1
·
verified ·
1 Parent(s): d0ae944

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -18,7 +18,7 @@ datasets:
18
 
19
  # tinydit-256
20
 
21
- A 210M-parameter text-to-image diffusion transformer trained from scratch on one GPU (3.5 days, 400k steps,
22
  4.2M images at 256² area with aspect-ratio buckets). Frozen FLUX.2 autoencoder (32-channel latent) and frozen
23
  flan-t5-base text encoder; the transformer, recipe, data pipeline and evaluation are original.
24
 
 
18
 
19
  # tinydit-256
20
 
21
+ A 210M-parameter text-to-image diffusion transformer trained from scratch on one GPU (RTX Pro 6000, 3.5 days, 400k steps,
22
  4.2M images at 256² area with aspect-ratio buckets). Frozen FLUX.2 autoencoder (32-channel latent) and frozen
23
  flan-t5-base text encoder; the transformer, recipe, data pipeline and evaluation are original.
24