Google批处理预测错误:运行图形期间的异常:断言失败：[无法将字节解码为JPEG、PNG、GIF或BMP]

Question

我在Google中部署了一个对象检测模型，我可以制作在线预测，但是失败了使批处理预测，堆栈驱动程序日志中的错误如下：

运行图形期间的异常:断言失败:无法将字节解码为JPEG、PNG、GIF或BMP[Node: map/while/decode_image/cond_jpeg/cond_png/cond_gif/Assert_1/Assert = Assertdevice="/job:localhost/replica:0/task:0/device:CPU:0"]]

我尝试了gcloud命令和python，但没有成功。用于在线预测的request.json文件。

{"inputs": {"b64": "/9j/4SurRXhpZgAATU0AKgAAAAgACgEPAAIAAAAHAAAAhgEQAAIAAAAFAAAAjgEaAAUAAAABAAAAlAEbAAUAAAABAAAAnAEoAAMAAAABAAIAAAExAAA2gITAAMAAAABAAEAAIdpAAQAAAABAAAA7oglAAQAAAABAAAC0gAAAyhYaWFvbWkAAE1.....}}

它已经被b64编码了。它适用于在线预测：

gcloud ml-engine predict --model object_detector  --version v2 --json-instances request.json

但是对于批处理预测失败，下面是batch_request.json文件的两行

{'instances': [{"inputs": {"b64": "/9j/4SurRXhpZgAATU0AKgAAAAgACgEPAAIAHAAAAhgEQAAIAAAAFAAAAjgEaAAUAAAABAAAAlAEbAAUAAAABAAAAnAEoAAMAAAABAAIAAAExAAIAAAA1AAAApAEyAAIAAAAUA...}}]}
{'instances': [{"inputs": {"b64": "/9j/4SurRXhpZgAATU0AKgAAAAgACgEPAAIAAAAAAhgEQAAIAAAAFAAAAjgEaAAUAAAABAAAAlAEbAAUAAAABAAAAnAEoAAMAAAABAAIAAAExAAIAAAA1AAAApAEyAAIAAAAUA...}}]}

用于批处理预测的python请求的主体：

{'jobId': 'mycloud_machine_object_detector_115252',
'predictionInput': {'dataFormat': 'TEXT',
'inputPaths': 'gs://my-bucket/object-detection/batch_request.json',
'outputPath': 'gs://my-bucket/object-detection/',
'region': 'us-central1',
'versionName': 'projects/mycloud_machine/models/object_detector/versions/v2'}}

我使用Google的python代码进行批处理请求。

project_id = 'projects/{}'.format(project_name)

ml = discovery.build('ml', 'v1', credentials=credentials)
request = ml.projects().jobs().create(parent=project_id,
                                  body=body_fn())

try:
    response = request.execute()

    print('Job requested.')

    # The state returned will almost always be QUEUED.
    print('state : {}'.format(response['state']))

except errors.HttpError as err:
    # Something went wrong, print out some information.
    print('There was an error getting the prediction results.' +
      'Check the details:')
    print(err._get_reason())

Answer 1

尝试此格式进行批处理预测：

{"inputs": {"b64": "/9j/4SurRXhpZgAATU0AKgAAAAgACgEPAAIAHAAAAhgEQAAIAAAAFAAAAjgEaAAUAAAABAAAAlAEbAAUAAAABAAAAnAEoAAMAAAABAAIAAAExAAIAAAA1AAAApAEyAAIAAAAUA...}}
{"inputs": {"b64": "/9j/4SurRXhpZgAATU0AKgAAAAgACgEPAAIAAAAAAhgEQAAIAAAAFAAAAjgEaAAUAAAABAAAAlAEbAAUAAAABAAAAnAEoAAMAAAABAAIAAAExAAIAAAA1AAAApAEyAAIAAAAUA...}}

换句话说，与gcloud一起使用的发送在线预测请求的格式相同。

批处理预测、gcloud和在线预测请求之间的关系如下：

• local predict和predict的文件格式与批处理预测相同
• 当使用gcloud时发送到联机预测的请求的实际主体将文件的每一行转换为“实例”数组的元素。因此，请求的实际主体(例如，如果您不使用gcloud，即curl、Python的request库等)将发送给在线预测服务。是{"instances": [line1, line2, ...]}