我使用SavedModel训练并导出了一个模型,正如Online Predict所建议的那样,我在本地进行了测试并且工作正常。
当我要求预测时,我没有解决,无论是在线预测还是批量预测。对于在线预测,我得到了:
jsonPayload: {
@type: "type.googleapis.com/google.cloud.ml.v1.PredictionLogEntry"
message: "{"error": "Prediction failed: Exception during model execution:
AbortionError(code=StatusCode.FAILED_PRECONDITION,
details=\"Attempting to use uninitialized value bout/weight\n\t
[[Node: bout/weight/read = Identity[T=DT_FLOAT,
_class=[\"loc:@bout/weight\"],
_output_shapes=[[1,2]],
_device=\"/job:localhost/replica:0/task:0/cpu:0\"](bout/weight)]]
\")"}"
numInstances: "1"
和批量预测:
16:21:35.855
PermanentException: Failed to load the model due to bad model data. [while running 'BATCH_PREDICTION/Prediction/ParDo(PredictionDoFn)/Do']
{
insertId: "xunb3og1phwaag"
logName: "projects/toycloudml/logs/ml.googleapis.com%2Fdnn_test3"
receiveTimestamp: "2017-07-11T19:21:35.855457468Z"
resource: {
labels: {
job_id: "dnn_test3"
project_id: "toycloudml"
task_name: "service"
}
type: "ml_job"
}
severity: "ERROR"
textPayload: "PermanentException: Failed to load the model due to bad model data. [while running 'BATCH_PREDICTION/Prediction/ParDo(PredictionDoFn)/Do']
"
timestamp: "2017-07-11T19:21:35.855457468Z"
}
我使用gcloud ml-engine models create $MODEL_NAME --regions=$REGION --enable-logging
创建了我的模型,
gcloud ml-engine versions create v1 --model $MODEL_NAME --origin $MODEL_BINARIES --runtime-version 1.2
的版本
并使用gcloud ml-engine jobs submit prediction $JOB_NAME --model $MODEL_NAME --version v1 --data-format TEXT --region $REGION --input-paths $INPUT_PATHS --output-path $OUTPUT_PATH
我无法弄清楚为什么会发生这种情况,因为每个地方预测工作得很好。我错过了什么吗?