Azure / Azure/MachineLearningNotebooks

Use `OutputFileDatasetConfig` instead of `PipelineData` for Hyperdrive's `metrics_output`

Đang mở
#1,616 0 bình luận 0 reaction 0 người được giao Xem trên GitHub
Ngôn ngữ chính
Jupyter Notebook
Star
4.4k
Fork
2.6k
Chỉ số merge pull request
Không có pull request nào được merge trong 30 ngày

Mô tả

`HyperDriveStep` can store the `metrics_output` into a datastore. However, this datastore location can only be specified with [`PipelineData`](https://docs.microsoft.com/en-us/python/api/azureml-pipeline-core/azureml.pipeline.core.pipelinedata?view=azure-ml-py), which is not the recommended way of moving data across pipeline steps (see [here](https://docs.microsoft.com/en-us/azure/machine-learning/how-to-move-data-in-out-of-pipelines)).

I would love if these ways of transferring data across steps were consistent. In the meantime, how can I specify the exact location where I want the JSON file from Hyperdrive to be stored? I would like to store it within some subfolders but I don't understand how to do that. Moreover, the file created and saved by the HD step does not have an extension (it has only the name defined in the `PipelineData`) and it is stored in a generic folder `azureml` within the container:

```python
hd_data = PipelineData(
"hd_metrics",
datastore=ds,
output_mode='upload'
)
```
will save the file `hd_metrics` (no extension) in `{datastore}/azureml/{run-id}/hd_metrics`. I'd like to have more control on where and how this file is stored.

---
#### Document Details

⚠ *Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.*

* ID: 74c38a69-084f-2185-60b2-b5bdedd45c65
* Version Independent ID: d9d1234e-c713-ba08-fb19-b9874a356a5a
* Content: [azureml.pipeline.steps.HyperDriveStep class - Azure Machine Learning Python](https://docs.microsoft.com/en-us/python/api/azureml-pipeline-steps/azureml.pipeline.steps.hyperdrivestep?view=azure-ml-py)
* Content Source: [AzureML-Docset/stable/docs-ref-autogen/azureml-pipeline-steps/azureml.pipeline.steps.HyperDriveStep.yml](https://github.com/MicrosoftDocs/MachineLearning-Python-pr/blob/live/AzureML-Docset/stable/docs-ref-autogen/azureml-pipeline-steps/azureml.pipeline.steps.HyperDriveStep.yml)
* Service: **machine-learning**
* Sub-service: **core**
* GitHub Login: @DebFro
* Microsoft Alias: **debfro**

Hướng dẫn đóng góp

Chưa lập chỉ mục được hướng dẫn đóng góp cho kho mã nguồn này

Đánh giá

Issue này chưa được đánh giá.

Nhận issue mới trong hộp thư của bạn

Bản tóm tắt ngắn những issue GitHub phù hợp với người mới.