windlx/url-classifier-model
TEXT GENERATIONConcurrency Cost:1Model Size:1.5BQuant:BF16Ctx Length:32kPublished:Feb 24, 2026License:mitArchitecture:Transformer0.0K Open Weights Cold

The windlx/url-classifier-model is a 1.5 billion parameter URL type classification model, fine-tuned by windlx using LoRA on the Qwen2.5-1.5B base model. It is specifically designed to classify URLs as either list pages or detail pages, achieving 99% accuracy on its validation set. This model is optimized for applications requiring automated URL categorization based solely on the URL string, such as SEO analysis, web crawling, and website analytics.

Loading preview...