bench-labs/PixelModel-v4

Text-to-Image • 40.1M • Updated 1 day ago • 16 • 5

Note a tiny latent diffusion transformer, and the result is a roughly 10x jump in FID.

bench-labs/PixelModel-v3

Text-to-Image • 919k • Updated 1 day ago • 90 • 7

Note New architecture, not a scale-up. SIREN decoder + FiLM + learned embeddings. Beats v1 FID.

bench-labs/PixelModel-v2

Text-to-Image • 200k • Updated 1 day ago • 54 • 9

Note scaled up

bench-labs/pixelmodel-v1

Text-to-Image • 23.7k • Updated 5 days ago • 106 • 10

Note x8.5 smaller, more efficient, better prompt understanding and larger training dataset

bench-labs/pixelmodel

Text-to-Image • 203k • Updated 9 days ago • 66 • 6

Note The first of its kind.