Question

我在Google ML中部署了一个对象检测模型，我可以进行在线预测，但 FAILS 进行批量预测，堆栈驱动程序日志中出现以下错误：

运行图形时出现异常：断言失败：[无法将字节解码为JPEG，PNG，GIF或BMP] [[节点：map / while / decode_image / cond_jpeg / cond_png / cond_gif / Assert_1 / Assert = Assert [T = [DT_STRING]，汇总= 3，_device =＆＃34; / job：localhost / replica：0 / task：0 / device：CPU：0＆＃34;]（map / while / decode_image / cond_jpeg / cond_png / cond_gif / is_bmp，map / while / decode_image / cond_jpeg / cond_png / cond_gif / Assert_1 / Assert / data_0）]]

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

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

它已经是b64编码的。它适用于在线预测：

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

但批量预测失败，下面是batch_request.json文件的两行

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

为批量预测做出的python api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文档中的python代码进行批量请求。

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

尝试使用此格式进行批量预测：

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

换句话说，与gcloud用于发送在线预测请求的格式相同。

批量预测，gcloud和在线预测请求之间的关系如下：

gcloud的local predict和predict文件格式与批量预测相同
使用gcloud时发送到在线预测的请求的实际主体将文件的每一行转换为“instances”数组的元素。所以请求的实际主体（例如，如果你没有使用gcloud，你会发送到在线预测服务的那个，即使用curl，Python的request库等）是{"instances": [line1, line2, ...]}

Google ML批量预测错误：运行图表时出现异常：断言失败：[无法将字节解码为JPEG，PNG，GIF或BMP]

1 个答案: