xuanwulab/HaS_Text_0209_0.6B
TEXT GENERATIONConcurrency Cost:1Model Size:0.8BQuant:BF16Ctx Length:32kPublished:Feb 10, 2026License:mitArchitecture:Transformer Open Weights Warm

The xuanwulab/HaS_Text_0209_0.6B is a 0.6 billion parameter, full-precision (FP16) on-device privacy model developed by xuanwulab. It provides a complete pipeline for entity recognition, anonymization, and reversible restoration, ensuring data privacy without leaving the device. This model supports 8 languages natively and is designed for local inference, making it suitable for privacy-sensitive applications requiring on-device processing.

Loading preview...