It's learning to draw pump.fun. Give it your coin.
A small image model trained from zero on nothing but pump.fun coin art, on one RTX 4090. Paste a CA. It blurs that coin's logo into noise, then draws it back with whatever it has learned so far.
Connecting to the trainer…
Watch it learn
64 fixed seeds. The same noise goes in every time and only the model changes. Drag back to checkpoint 0 to see where it started.
Claim a seed
Burn 100,000 $DENOISE and you get the next numbered seed. A seed is a fixed block of noise tied to your wallet. At every checkpoint the model draws it again from that same noise, so you watch your seed turn from static into a character. Each seed gets a page with its whole timelapse and a share card.
The burn is one SPL burn instruction from your own wallet. The site checks it on-chain before it gives out a seed, and each burn claims one seed.
Already burned? Paste the transaction signature
Seed pool
Bigger runs
Each run starts again from noise at a higher resolution. Trading fees unlock the next one. The weights get published at each milestone.
Compute ledger
Real spend only. Rented compute gets a line here with the date, the hardware and the cost.
Live trades
No coin yet. Buys and sells show here once the CA is live.
What runs where
The model. A UNet diffusion model, about 20M parameters for the 32 px run, trained from random weights. No pretrained model and no other dataset is involved.
The data. pump.fun coin images only, pulled from pump.fun's own feeds while it trains, with coins launched in the last few hours and coins trending right now shown to it more often. NSFW images, hate symbols, re-uploads, brand copycats and wash "fund" coins are dropped (open-source image classifiers help pick those out; they never touch the model).
The pictures. Every tile, seed frame and redraw on this site is drawn by this model alone: the checkpoint named on it, sharpened by one of its own earlier checkpoints (a method called autoguidance). No other model draws anything. When the trainer is off, the page says so and shows the last checkpoint.
Seeds. Seed N always starts from the same noise, so its timelapse shows the model changing and nothing else. Seeds claimed later are also drawn through the earlier checkpoints of the run.