A decent text-to-image model trained for ~$20K (Apache 2.0 Weights)
Comments
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Summary
A two-person research lab released Pyramid-JiT, a 2B-parameter pixel-space text-to-image model with Apache 2.0 weights, saying its full training run cost about $21K on 64 H100s. The architecture compresses images with 32×32 patches instead of relying on a separate VAE, cutting attention context fourfold.