lq@lq:~/Fine-tuning$ cp backend/scripts/test_ms_api.py backend/data/ && sudo docker exec -it finetune-backend python3 /root/Fine-tuning/backend/data/test_ms_api.py tany0699/carBrands50
测试数据集: tany0699/carBrands50

=== Test1: 数据集 info API 完整响应 ===
请求: https://www.modelscope.cn/api/v1/datasets/tany0699/carBrands50
{
  "RequestId": "ec91c118-ce21-47eb-a861-8663b5483315",
  "Code": 200,
  "Message": "success",
  "Data": {
    "isTop": null,
    "relatedPaperId": null,
    "nexa": null,
    "Id": 2119,
    "Namespace": "tany0699",
    "Name": "carBrands50",
    "CreatedBy": "tany0699",
    "ChineseName": "车型分类",
    "License": "Apache License 2.0",
    "Description": "50种汽车类型分类数据集",
    "Visibility": 3,
    "Type": 2,
    "Owner": "tany0699",
    "UserDefineTags": "car,car brands,classification",
    "Likes": 19,
    "Downloads": 7664,
    "Size": null,
    "ReadmeContent": "\n## 数据集描述\n### 数据集简介\n数据集包含50种类型汽车的图像，其中训练集4260张图片，验证集99张图片，数据大小共46MB，支持识别车型包括：BMW、Audi、Jeep、Mini、Suzuki、Bentley等汽车品牌。  \n<img src=\"./sample.jpg\" alt=\"数据示例\"/>  \n\n### 数据集支持的任务\n可用于快速模型验证、性能评估、小数据分类集训练等。\n\n## 数据集的格式和结构\n### 数据格式\n数据集包括训练集train和验证集val，train和val文件夹之下按文件夹进行分类，共有2个子文件夹，同类别标签的图片在同一个文件夹下，图片格式为JPG。同时包含与标注文件中label id相对应的类名文件classname.txt。\n\n### 数据集加载方式\n```python\nfrom modelscope.msdatasets import MsDataset\nfrom modelscope.utils.constant import DownloadMode\n\nms_train_dataset = MsDataset.load(\n            'carBrands50', namespace='tany0699',\n            subset_name='default', split='train') # 加载训练集\nprint(next(iter(ms_train_dataset)))\n\nms_val_dataset = MsDataset.load(\n            'carBrands50', namespace='tany0699',\n            subset_name='default', split='validation') # 加载验证集\nprint(next(iter(ms_val_dataset)))\n```\n### 数据分片\n本数据集包含train和val数据集。\n| 子数据集    |        train | val |     test |\n|---------|-------------:|-----------:|---------:|\n| default |  训练集 |  验证集  | / |\n\n## 数据集生成的相关信息\n### 原始数据\nCar Brands Images：https://www.kaggle.com/datasets/yamaerenay/100-images-of-top-50-car-brands\n\n### Clone with HTTP\n```bash\ngit clone https://www.modelscope.cn/datasets/tany0699/carBrands50.git\n```",
    "AlreadyStar": false,
    "GmtCreate": 1675332057,
    "GmtModified": 1779949098,
    "Tags": [
      {
        "id": 3681,
        "datasetId": 2119,
        "domain": "image",
        "task": "图像分类",
        "filter": "样本规模",
        "label": "100-10k",
        "dataType": 2,
        "level1Tag": "图像分类",
        "level1TagName": "image-classification",
        "level2Tag": "样本规模",
        "level2TagName": "size_scale",
        "level3Tag": "100-10k",
        "level3TagName": "100-10k",
        "gmtCreate": 1675426534,
        "gmtModified": 1675426534,
        "tagId": null
      }
    ],
    "Status": 1,
    "FullName": null,
    "Organization": null,
    "UsedFor": null,
    "CertificationMark": 0,
    "LastUpdatedTime": 1675426534,
    "FromSite": "maas",
    "SourcePlatform": null,
    "topIndex": null,
    "IsFlex": 0,
    "StorageSize": 220619,
    "RelateArxivId": null,
    "Avatar": null,
    "ProtectedMode": 2,
    "ApprovalMode": null,
    "ApprovalNotifyEmail": null,
    "ApplyMeta": null,
    "NEXA": null
  },
  "PageNumber": null,
  "PageSize": null,
  "TotalCount": null
}

=== Test2: HubApi 直接调用 ===
尝试 get_dataset_files...
get_dataset_files 失败: 'HubApi' object has no attribute 'get_dataset_files'

尝试 get_dataset_file_url...
train.csv 下载 URL: https://www.modelscope.cn/api/v1/datasets/tany0699/carBrands50/repo?Source=SDK&Revision=master&FilePath=train.csv&View=False

=== Test3: carBrands50.json 配置文件 ===
请求: https://www.modelscope.cn/api/v1/datasets/tany0699/carBrands50/repo?Revision=master&FilePath=carBrands50.json&View=false
{"default":{"train":{"meta":"train.csv","file":"train.zip"},"test":{"meta":"","file":""},"validation":{"meta":"val.csv","file":"val.zip"}}}

=== Test4: dataset_infos.json ===
请求: https://www.modelscope.cn/api/v1/datasets/tany0699/carBrands50/repo?Revision=master&FilePath=dataset_infos.json&View=false
{"default":{"features":{"image":{"_type":"Image"},"category":{"_type":"Value"}},"splits":{"validation":{"name":"validation","dataset_name":"carBrands50"},"train":{"name":"train","dataset_name":"carBrands50"}}}}
