WolleWeb-1.7B

Fine-tuned from Qwen3-1.7B for local, private, in-browser AI.

Trained to reduce hallucinations by returning search queries instead of guessing (e.g. <search>query inside</search>), specialized on English and Chinese, and optimized for natural conversation.

Part of the Wolle.AI project, built to make local AI accessible to everyone.

Try it at WolleWeb.

Trained by gr

Downloads last month: 8

Safetensors

Model size

2B params

Tensor type

BF16

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for wolledotai/Wolleweb-1.7B-v0

Quantizations

2 models