---
title: ViSpeak-s2
canonical_url: "https://www.modelscope.cn/models/fushh7/ViSpeak-s2"
md_url: "https://www.modelscope.cn/models/fushh7/ViSpeak-s2.md"
repository: fushh7/ViSpeak-s2
last_updated: 2025-03-18
pipeline_tag: image-text-to-text
tasks:
  - image-text-to-text
model_type:
  - vita-Qwen2
architectures:
  - VITAQwen2ForCausalLM
parameters: 8.4B
tensor_type:
  - F16
  - F32
library_name:
  - safetensors
  - pytorch
frameworks:
  - Pytorch
downloads: 28
stars: 0
---

# ViSpeak-s2

> ViSpeak-s2 - fushh7 在 ModelScope 开源的模型。This repository contains the model of the paper ViSpeak: Visual Instruction Feedback in Streaming Videos.

- **Repository**: fushh7/ViSpeak-s2
- **Tasks**: image-text-to-text
- **Parameters**: 8.4B
- **Downloads**: 28
- **Stars**: 0
- **Last updated**: 2025-03-18

Source: https://www.modelscope.cn/models/fushh7/ViSpeak-s2

---

This repository contains the model of the paper [ViSpeak: Visual Instruction Feedback in Streaming Videos](https://arxiv.org/abs/2503.12769).

Code: https://github.com/HumanMLLM/ViSpeak