Fine-tuning a Jev-style classifier for ~$1.50 lifts tag accuracy from 10% to 69%
A 194M GLiNER2 model fine-tuned on Hugging Face Jobs for about $1.50 suggests task tags for any Hub dataset from its column names and first row, raising the match rate from 10% zero-shot to 69%.
Trained a Jev-style classifier on @huggingface Jobs for ~$1.50. It's a 194M GLiNER2 model that suggests task tags for any Hub dataset from its column names and first row, and returns a label with a probability. Zero-shot, GLiNER2's first suggestion matched an owner's tag 10% of the time. After 17 minutes of fine-tuning: 69%. The fine-tuned model runs on a free CPU in about a second. Owners' tags are noisy, so some "wrong" answers are tags the owner left out. The recipe is open: one hf jobs command trains the same kind of model on your own labels. The README example (book titles) runs in ~2 minutes for about $0.02. Demo: http://huggingface.co/spaces/davanstrien/hub-task-tagger Recipe: http://huggingface.co/datasets/uv-scripts/classification