TabPFN and TabICL against tuned XGBoost: the model that does not train won on fourteen tables out of fourteen
The claim behind TabPFN and TabICL: they predict on a table without ever training on it, and still beat tuned boosting. Measured on 14 datasets from the Grinsztajn benchmark, same split and same clock for everyone. The model that does not train won on 14 out of 14 against tuned XGBoost. What it costs in latency and VRAM, and when I'd still reach for boosting — in the post. Read the full…
We haven't written up this one. Dev.to has the full story — the link below goes straight to it.
