Caption_With_Modality
收藏魔搭社区2025-12-02 更新2025-11-15 收录
下载链接:
https://modelscope.cn/datasets/SongtaoJ/Caption_With_Modality
下载链接
链接失效反馈官方服务:
资源简介:
# Hulu-Med Dataset
### A Transparent Generalist Model towards Holistic Medical Vision-Language Understanding
This repository provides the dataset used in the paper:
**Hulu-Med: A Transparent Generalist Model towards Holistic Medical Vision-Language Understanding**
The dataset is released in **three stages**, each split into multiple `.tar` part files.
To use the dataset, simply **merge the parts** and **extract** them.
---
## 0. Download (ModelScope)
```bash
modelscope download --dataset SongtaoJ/Caption_With_Modality --local_dir ./hulumed_data
```
---
## 1. Dataset Structure
```
hulumed_data/
├── Stage0/
│ ├── hulu_med_stage0_part00.tar
│ ├── hulu_med_stage0_part01.tar
│ ├── hulu_med_stage0_part02.tar
│ └── ...
├── Stage1/
│ ├── hulu_med_stage1_part00.tar
│ └── ...
└── Stage2/
├── hulu_med_stage2_part00.tar
└── ...
```
---
## 2. Merge & Extract
### Stage0
```bash
cd hulumed_data/Stage0
cat hulu_med_stage0_part*.tar > hulu_med_stage0.tar
tar -xf hulu_med_stage0.tar -C /path/to/output/Stage0
```
### Stage1
```bash
cd hulumed_data/Stage1
cat hulu_med_stage1_part*.tar > hulu_med_stage1.tar
tar -xf hulu_med_stage1.tar -C /path/to/output/Stage1
```
### Stage2
```bash
cd hulumed_data/Stage2
cat hulu_med_stage2_part*.tar > hulu_med_stage2.tar
tar -xf hulu_med_stage2.tar -C /path/to/output/Stage2
```
---
## 3. Final Output Layout
```
/path/to/output/
├── Stage0/
├── Stage1/
└── Stage2/
```
# Hulu-Med 数据集
### 面向全方位医学视觉语言理解的透明通用模型(Transparent Generalist Model towards Holistic Medical Vision-Language Understanding)
本仓库提供了论文 **《Hulu-Med:面向全方位医学视觉语言理解的透明通用模型》** 中使用的数据集。
该数据集以**三阶段**形式发布,每个阶段拆分为多个`.tar`分卷文件。
使用该数据集时,仅需先**合并分卷**再**解压**即可。
---
## 0. 下载(ModelScope)
bash
modelscope download --dataset SongtaoJ/Caption_With_Modality --local_dir ./hulumed_data
---
## 1. 数据集结构
hulumed_data/
├── Stage0/
│ ├── hulu_med_stage0_part00.tar
│ ├── hulu_med_stage0_part01.tar
│ ├── hulu_med_stage0_part02.tar
│ └── ...
├── Stage1/
│ ├── hulu_med_stage1_part00.tar
│ └── ...
└── Stage2/
├── hulu_med_stage2_part00.tar
└── ...
---
## 2. 合并与解压
### Stage0
bash
cd hulumed_data/Stage0
cat hulu_med_stage0_part*.tar > hulu_med_stage0.tar
tar -xf hulu_med_stage0.tar -C /path/to/output/Stage0
### Stage1
bash
cd hulumed_data/Stage1
cat hulu_med_stage1_part*.tar > hulu_med_stage1.tar
tar -xf hulu_med_stage1.tar -C /path/to/output/Stage1
### Stage2
bash
cd hulumed_data/Stage2
cat hulu_med_stage2_part*.tar > hulu_med_stage2.tar
tar -xf hulu_med_stage2.tar -C /path/to/output/Stage2
---
## 3. 最终输出目录结构
/path/to/output/
├── Stage0/
├── Stage1/
└── Stage2/
提供机构:
maas
创建时间:
2025-07-01



