add coco verification how-to

Author: PatrickFarley

articles/cognitive-services/Computer-vision/how-to/coco-verification.md

@@ -0,0 +1,49 @@
+---
+title: Verify a COCO annotation file
+titleSuffix: Azure Cognitive Services
+description: Use a Python script to verify your COCO file for custom model training.
+services: cognitive-services
+author: PatrickFarley
+manager: nitinme
+ms.service: cognitive-services
+ms.topic: how-to
+ms.date: 03/21/2023
+ms.author: pafarley
+---
+
+<!-- nbstart https://raw.githubusercontent.com/Azure-Samples/cognitive-service-vision-model-customization-python-samples/main/docs/check_coco_annotation.ipynb -->
+
+> [!TIP]
+> Contents of _check_coco_annotation.ipynb_. **[Open in GitHub](https://github.com/Azure-Samples/cognitive-service-vision-model-customization-python-samples/blob/main/docs/check_coco_annotation.ipynb)**.
+
+# Check the format of your COCO annotation file
+
+This notebook demonstrates how to check if the format of your annotation file is correct. First, install the python samples package from the command line:
+
+```python
+pip install cognitive-service-vision-model-customization-python-samples
+```
+
+Then, run the following python code to check the file's format. You can either enter this code in a Python script, or run the [Jupyter Notebook](https://github.com/Azure-Samples/cognitive-service-vision-model-customization-python-samples/blob/main/docs/check_coco_annotation.ipynb) on a compatible platform.
+
+```python
+from cognitive_service_vision_model_customization_python_samples import check_coco_annotation_file, AnnotationKind, Purpose
+import pathlib
+import json
+
+coco_file_path = pathlib.Path("{your_coco_file_path}")
+annotation_kind = AnnotationKind.MULTICLASS_CLASSIFICATION # or AnnotationKind.OBJECT_DETECTION
+purpose = Purpose.TRAINING # or Purpose.EVALUATION
+
+check_coco_annotation_file(json.loads(coco_file_path.read_text()), annotation_kind, purpose)
+```
+
+<!-- nbend -->
+
+## Use COCO file in a new project
+
+Once your COCO file is verified, you're ready to import it to your model customization project. See [Create and train a custom model](model-customization.md) and go to the section on selecting a COCO file&mdash;you can follow the guide from there to the end.
+
+## Next steps
+
+* [Create and train a custom model](model-customization.md)

articles/cognitive-services/Computer-vision/how-to/model-customization.md

@@ -88,7 +88,7 @@ Note that
 
 - Image Analysis model customization for classification is different from other kinds of vision training, as we utilize your class names, as well as image data, in training. So, be sure provide meaningful category names in the annotations.
 - Note that in the example dataset, there are few images for the sake of simplicity. Although [Florence models](https://www.microsoft.com/en-us/research/publication/florence-a-new-foundation-model-for-computer-vision/) achieve great few-shot performance (high model quality even with little data available), it's good to have more data for the model to learn. Our recommendation is to have at least five images per class, and the more the better.
-- Once your COCO annotation file is prepared, you can use the [COCO file verification script](tbd) to check the format.
+- Once your COCO annotation file is prepared, you can use the [COCO file verification script](coco-verification.md) to check the format.
 
 #### Multiclass classification example
 

articles/cognitive-services/Computer-vision/includes/coco-files.md

@@ -118,4 +118,8 @@ The value for `absolute_url` can be found in your blob container's properties:
 | Key | Type | Description | Required? |
 |-|-|-|-|
 | `id` | integer | Unique ID for each category (label class). These should be present in the `annotations` section. | Yes |
-| `name`| string | Name of the category (label class) | Yes |
+| `name`| string | Name of the category (label class) | Yes |
+
+### COCO file verification
+
+You can use our [Python sample code](/azure/cognitive-services/computer-vision/how-to/coco-verification) to check the format of a COCO file.

articles/cognitive-services/Computer-vision/toc.yml

@@ -146,6 +146,8 @@ items:
         href: how-to/model-customization.md
       - name: Migrate a Custom Vision project to Image Analysis
         href: how-to/migrate-from-custom-vision.md
+      - name: Verify a custom model COCO file
+        href: how-to/coco-verification.md
       - name: Call the Image retrieval APIs
         href: how-to/image-retrieval.md
       - name: Call the Background removal API