datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
clip-guided-diffusion-aoti-zerogpuvqgan-clip-aoti-zerogpuiab-url-gold-784
IAB URL Gold 784
A 784-row gold benchmark for URL-only IAB classification: given nothing but a bare
domain (rendered as a short text string), predict the
IAB Tech Lab content and audience
categories of the web destination.
This is the held-out evaluation set used to benchmark
ZeroGPU/zlm-v1-iab-domain-classifier
against GPT-5.4-nano. On it, the fine-tuned 149M ModernBERT model reaches content
micro-F1 0.3845 (P 0.4247 / R 0.3512) vs the nano baseline's 0.3526
(P 0.3332 / R… See the full description on the dataset page: https://huggingface.co/datasets/ZeroGPU/iab-url-gold-784.