使用Python SDK向Custom Vision批量上传带标签图片报错求助

错误原因 #

你遇到的'Image' object has no attribute 'error'错误，核心是对Custom Vision SDK返回结果的属性访问路径错误。upload_result.images中的每个元素是ImageFileCreateResult对象，它的错误信息、图片ID等属性直接挂载在对象本身，而非嵌套在image子属性中。

代码修复 #

修改批量上传后的结果处理逻辑，调整属性访问路径：

if not upload_result.is_batch_successful:
    for image in upload_result.images:
        if image.status != "OK":
            print(
                f"Image '{image.name}' upload failed. Status: {image.status}, Error Details: {image.error.message if image.error else 'Unknown'}"
            )
        else:
            print(
                f"Image '{image.name}' uploaded successfully with ID: {image.id}")

额外优化建议 #

• 处理剩余图片：当前代码仅在image_batch达到设定的batch_size时执行上传，循环结束后会遗漏最后一批不足batch_size的图片，建议在循环结束后补充以下代码：

  # 上传最后一批剩余图片
  if image_batch:
      try:
          upload_result = trainer.create_images_from_files(
              PROJECT_ID, ImageFileCreateBatch(images=image_batch))
          if not upload_result.is_batch_successful:
              for image in upload_result.images:
                  if image.status != "OK":
                      print(
                          f"Image '{image.name}' upload failed. Status: {image.status}, Error Details: {image.error.message if image.error else 'Unknown'}"
                      )
                  else:
                      print(
                          f"Image '{image.name}' uploaded successfully with ID: {image.id}")
      except Exception as e:
          print(f"Final batch upload failed: {e}")
  

• 优化标签获取逻辑：当前每次循环都调用trainer.get_tag请求API获取标签ID，存在重复请求的冗余。建议提前一次性获取所有标签并缓存为字典：

  # 预获取所有标签并缓存
  tags = trainer.get_tags(PROJECT_ID)
  tag_id_map = {tag.name: tag.id for tag in tags}
  
  # 后续处理标签时直接从字典读取
  for key, value in row.items():
      if key != "filename" and value == "1":
          if key in tag_id_map:
              tag_ids.append(tag_id_map[key])
          else:
              print(f"Warning: Tag '{key}' not found in project")
  

内容的提问来源于stack exchange，提问作者sweecodengo