---
title: ChatTruth-7B
canonical_url: "https://www.modelscope.cn/models/mirror013/ChatTruth-7B"
md_url: "https://www.modelscope.cn/models/mirror013/ChatTruth-7B.md"
repository: mirror013/ChatTruth-7B
last_updated: 2024-03-15
pipeline_tag: image-classification
tasks:
  - image-classification
model_type:
  - qwen
architectures:
  - QWenLMHeadModel
library_name:
  - pytorch
  - transformer
frameworks:
  - Pytorch
language:
  - zh
  - en
inference_backends:
  - "deploy_task vlm/text/emb"
  - "lmdeploy 0.9.1"
  - "lmdeploy_turbomind 0.9.1"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 109
stars: 0
---

# ChatTruth-7B

> ChatTruth-7B - mirror013 在 ModelScope 开源的模型。ChatTruth-7B 在Qwen-VL的基础上，使用精心设计的数据进行了优化训练。与Qwen-VL相比，模型在大分辨率上得到了大幅提升。创新性提出Restore Module使大分辨率计算量大幅减少。

mirror013/ChatTruth-7B 是 ModelScope 魔搭社区上的image-classification模型，可用 deploy_task vlm/text/emb、lmdeploy 0.9.1、lmdeploy_turbomind 0.9.1 部署。

- **Repository**: mirror013/ChatTruth-7B
- **Tasks**: image-classification
- **Inference backends**: deploy_task vlm/text/emb, lmdeploy 0.9.1, lmdeploy_turbomind 0.9.1, sglang 0.5.2, vllm 0.9.2
- **Downloads**: 109
- **Stars**: 0
- **Last updated**: 2024-03-15

Source: https://www.modelscope.cn/models/mirror013/ChatTruth-7B

---

# ChatTruth-7B

**ChatTruth-7B** 在Qwen-VL的基础上，使用精心设计的数据进行了优化训练。与Qwen-VL相比，模型在大分辨率上得到了大幅提升。创新性提出Restore Module使大分辨率计算量大幅减少。

![image/png](https://cdn-uploads.huggingface.co/production/uploads/657bef8a5c6f0b1f36fcf28e/kwgU2AxZbJzxmgWULwv6A.png)

## 安装要求 (Requirements)

* transformers 4.32.0
* python 3.8 and above
* pytorch 1.13 and above
* CUDA 11.4 and above
  
  <br>

## 快速开始 (Quickstart)

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
from transformers.generation import GenerationConfig
import torch
torch.manual_seed(1234)
model_path = 'ChatTruth-7B' # your downloaded model path.

tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)

# use cuda device
model = AutoModelForCausalLM.from_pretrained(model_path, device_map="cuda", trust_remote_code=True).eval()

model.generation_config = GenerationConfig.from_pretrained(model_path, trust_remote_code=True)
model.generation_config.top_p = 0.01

query = tokenizer.from_list_format([
    {'image': 'demo.jpeg'},
    {'text': '图片中的文字是什么'},
])
response, history = model.chat(tokenizer, query=query, history=None)
print(response)

# 昆明太厉害了
```