view2space-4b / README.md
multimodalart's picture
multimodalart HF Staff
Upload folder using huggingface_hub
9ead0e0 verified
|
Raw
History Blame Contribute Delete
994 Bytes

A newer version of the Gradio SDK is available: 6.20.0

Upgrade
metadata
title: VIEW2SPACE-4B
emoji: 👁
colorFrom: pink
colorTo: purple
sdk: gradio
sdk_version: 6.19.0
app_file: app.py
short_description: Multi-view visual reasoning VLM based on Qwen3-VL 4B
python_version: '3.12'
startup_duration_timeout: 30m

VIEW2SPACE-4B Demo

A Gradio demo for Pokerme/view2space_4b, a 4.4B-parameter vision-language model based on Qwen3-VL that performs multi-view visual reasoning from sparse observations.

Usage

Upload one or more images and type a question. The model integrates information across all provided views to produce an answer. Examples include counting objects, describing scenes, detecting regions, and answering multiple-choice spatial reasoning questions.

Reference