87 строки
3.0 KiB
Markdown
87 строки
3.0 KiB
Markdown
# Batch Scoring on Azure for Deep Learning Models
|
||
|
||
This tutorial demonstrates how to deploy a deep learning model to Azure for batch scoring.
|
||
|
||
### File System
|
||
```
|
||
.
|
||
├── bait/
|
||
│ ├── cluster_setup.py
|
||
│ ├── config.py
|
||
│ ├── config_template.py
|
||
│ ├── fileshare_setup.py
|
||
│ ├── job_setup.py
|
||
├── scoring_script/
|
||
│ ├── pytorch_classification/
|
||
│ │ └── score0.py
|
||
│ └── tf_mnist/
|
||
│ └── score0.py
|
||
├── training_script/
|
||
| ├── pytorch_classification/
|
||
| │ ├── train0.ipynb
|
||
| │ └── train0.py
|
||
| └── tf_mnist/
|
||
| ├── train0.ipynb
|
||
| └── train0.py
|
||
|
||
--- Files below this point will be generated in the tutorial
|
||
|
||
├── data/
|
||
│ └── pytorch_classification/
|
||
├── model/
|
||
│ ├── pytorch_classification/
|
||
│ └── tf_mnist/
|
||
└── func/
|
||
├── blobtrig/
|
||
│ ├── function.json
|
||
│ ├── host.json
|
||
│ ├── __init__.py
|
||
│ ├── readme.md
|
||
│ └── sample.dat
|
||
├── host.json
|
||
├── local.settings.json
|
||
└── requirements.txt
|
||
```
|
||
There are a few main folders to take note of in this repository:
|
||
|
||
__/bait__
|
||
|
||
This folder contains all the Batch AI scripts, including:
|
||
- `cluster_setup.py` - executed locally
|
||
- `fileshare_setup.py` - executed locally
|
||
- `job_setup.py` - executed by Functions V2
|
||
|
||
It also contains a `config_template.py` file, which needs to be renamed as `config.py` and filled out.
|
||
|
||
|
||
__/func__
|
||
|
||
This folder contains everything needed to run your functions v2. (TODO - maybe this should be made by the user?)
|
||
|
||
__/models__
|
||
|
||
This folder is where we store the model files that we will use for scoring.
|
||
|
||
__/scoring_script__
|
||
|
||
This folder contains the scoring script that will use a model in the /models directory. This scoring script will be executed on nodes in the Batch AI cluster.
|
||
|
||
__/training_script__
|
||
|
||
This folder contains the training scripts used to generate the models in the /models directory. This training script will be executed locally on a GPU enabled VM.
|
||
|
||
|
||
# Contributing
|
||
|
||
This project welcomes contributions and suggestions. Most contributions require you to agree to a
|
||
Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us
|
||
the rights to use your contribution. For details, visit https://cla.microsoft.com.
|
||
|
||
When you submit a pull request, a CLA-bot will automatically determine whether you need to provide
|
||
a CLA and decorate the PR appropriately (e.g., label, comment). Simply follow the instructions
|
||
provided by the bot. You will only need to do this once across all repos using our CLA.
|
||
|
||
This project has adopted the [Microsoft Open Source Code of Conduct](https://opensource.microsoft.com/codeofconduct/).
|
||
For more information see the [Code of Conduct FAQ](https://opensource.microsoft.com/codeofconduct/faq/) or
|
||
contact [opencode@microsoft.com](mailto:opencode@microsoft.com) with any additional questions or comments.
|