Posts
UNO: One Change to FLUX, and a Data Flywheel
Reading notes on UNO: how a minimal positional change, synthetic training pairs, and a less-to-more curriculum turn FLUX into a multi-subject customization model.
Generative ModelsSeedDiT Control Is a Few Decisions About Tokens
A working skeleton: what a diffusion transformer actually is, why every control method decomposes into a handful of token-level decisions, and a starter table mapping papers onto those axes.
Generative ModelsSeedHarness, Not Agent: Reading Notes on Lilian Weng's Harness Post
What a harness actually is, how I would pick harness components for my own research loop, and the failure patterns I recognize from my own training runs.
Agentic SystemsSeedA Garden, Not a Blog
Why this site publishes unfinished notes on purpose — the three tracks, the status system, and the ADHD-proof pipeline behind it.
Agentic SystemsEvergreenEssayOminiControl2: What It Proves, and What It Doesn't
Reading notes on efficient conditioning for diffusion transformers — where the multi-condition cost comes from, how the paper measures efficiency, and the gap it leaves in control fidelity evaluation.
Generative ModelsGrowing