Azure ML 实验批处理 Web 服务调用因无效输出扩展而失败

Question

我有一个 Azure webjob，它通过 HttpRequests 调用 ML 训练实验，利用 ML webportal 中生成的代码：

var request = new BatchExecutionRequest()
            {
                Inputs = new Dictionary<string, AzureBlobDataReference>() {
                    {
                        "input1",
                        new AzureBlobDataReference()
                        {
                            ConnectionString = _connectionString,
                            RelativeLocation = $"{_containerName}/{experimentId}/{tenantId}/{trainingDataFileName}"
                        }
                    },
                },

                Outputs = new Dictionary<string, AzureBlobDataReference>() {
                    {
                        "output1",
                        new AzureBlobDataReference()
                        {
                            ConnectionString = "azureStorageConnectionString",
                            RelativeLocation = $"{_containerName}/{experimentId}/{tenantId}/Model_2018421.ilearner"
                        }
                    },
                },

                GlobalParameters = new Dictionary<string, string>()
                {
                }
            };

但是，请求失败并显示以下消息：

The blob reference: experiments/experimentId/TenantId/Model_2018421.ilearner has an invalid or missing file extension. Supported file extensions for this output type are: \".csv, .tsv, .arff\"

我对此很困惑，因为如果我希望经过训练的模型使用“.ilearner”作为模型的文件扩展名，那么文档中到处都是这样写的。

我看过this question asking about the same error leveraging the DataFactory, and also this question on datascience.stackexchange。两人都没有任何线索、答案或其他后续行动。

任何关于我遗漏的见解都将不胜感激！

Answer 1

任何正在寻找你的 "Don't Overthink It" 时刻的人：

我需要提供两个输出 blob 文件引用：

var request = new BatchExecutionRequest()
            {
                Inputs = new Dictionary<string, AzureBlobDataReference>() {
                    {
                        "input1",
                        new AzureBlobDataReference()
                        {
                            ConnectionString = _connectionString,
                            RelativeLocation = $"{_containerName}/{experimentId}/{tenantId}/{trainingDataFileName}.csv"
                        }
                    },
                },

                Outputs = new Dictionary<string, AzureBlobDataReference>() {
                    {
                        "output1",
                        new AzureBlobDataReference()
                        {
                            ConnectionString = _connectionString,
                            RelativeLocation = $"{_containerName}/{experimentId}/{tenantId}/{outputFileNameCsv}.csv"
                        }
                    },
                    {
                        "output2",
                        new AzureBlobDataReference()
                        {
                            ConnectionString = _connectionString,
                            RelativeLocation = $"{_containerName}/{experimentId}/{tenantId}/{outputFileNameIlearner}.ilearner"
                        }
                    },
                },

                GlobalParameters = new Dictionary<string, string>()
                {
                }
            };

美式英语中有一句老话叫不做假设，我假设第二个输出是批处理操作中使用的可选参数。由于我实际上并没有从每次调用中寻找一个以上的结果，所以我认为删除第二个输出参数是安全的。

TL/DR：保留网络服务门户 "Consume" 选项卡生成的所有参数，并确保第一个参数是 .csv 文件引用。

Azure ML 实验批处理 Web 服务调用因无效输出扩展而失败

Azure ML Experiment Batch Webservice Call Fails with Invalid Output Extension

c#

rest

azure

azure-machine-learning-studio