run to trials,jobs

Author: shubham soni

articles/machine-learning/how-to-auto-train-image-models.md

@@ -23,7 +23,7 @@ ms.date: 07/13/2022
 
 In this article, you learn how to train computer vision models on image data with automated ML with the Azure Machine Learning CLI extension v2 or the Azure Machine Learning Python SDK v2.
 
-Automated ML supports model training for computer vision tasks like image classification, object detection, and instance segmentation. Authoring AutoML models for computer vision tasks is currently supported via the Azure Machine Learning Python SDK. The resulting experimentation runs, models, and outputs are accessible from the Azure Machine Learning studio UI. [Learn more about automated ml for computer vision tasks on image data](concept-automated-ml.md).
+Automated ML supports model training for computer vision tasks like image classification, object detection, and instance segmentation. Authoring AutoML models for computer vision tasks is currently supported via the Azure Machine Learning Python SDK. The resulting experimentation trials, models, and outputs are accessible from the Azure Machine Learning studio UI. [Learn more about automated ml for computer vision tasks on image data](concept-automated-ml.md).
 
 ## Prerequisites
 
@@ -102,7 +102,7 @@ In order to generate computer vision models, you need to bring labeled image dat
 If your training data is in a different format (like, pascal VOC or COCO), you can apply the helper scripts included with the sample notebooks to convert the data to JSONL. Learn more about how to [prepare data for computer vision tasks with automated ML](how-to-prepare-datasets-for-automl-images.md). 
 
 > [!Note]
-> The training data needs to have at least 10 images in order to be able to submit an AutoML run. 
+> The training data needs to have at least 10 images in order to be able to submit an AutoML job. 
 
 > [!Warning]
 > Creation of `MLTable` from data in JSONL format is supported using the SDK and CLI only, for this capability. Creating the `MLTable` via UI is not supported at this time. 
@@ -277,27 +277,27 @@ image_object_detection_job = automl.image_object_detection(
 
 ## Configure experiments
 
-For computer vision tasks, you can launch either [individual runs](#individual-runs), [manual sweeps](#manually-sweeping-model-hyperparameters) or [automatic sweeps](#automatically-sweeping-model-hyperparameters-automode). We recommend starting with an automatic sweep to get a first baseline model. Then, you can try out individual runs with certain models and hyperparameter configurations. Finally, with manual sweeps you can explore multiple hyperparameter values near the more promising models and hyperparameter configurations. This three step workflow (automatic sweep, individual runs, manual sweeps) avoids searching the entirety of the hyperparameter space, which grows exponentially in the number of hyperparameters.
+For computer vision tasks, you can launch either [individual trials](#individual-trials), [manual sweeps](#manually-sweeping-model-hyperparameters) or [automatic sweeps](#automatically-sweeping-model-hyperparameters-automode). We recommend starting with an automatic sweep to get a first baseline model. Then, you can try out individual trials with certain models and hyperparameter configurations. Finally, with manual sweeps you can explore multiple hyperparameter values near the more promising models and hyperparameter configurations. This three step workflow (automatic sweep, individual trials, manual sweeps) avoids searching the entirety of the hyperparameter space, which grows exponentially in the number of hyperparameters.
 
 Automatic sweeps can yield competitive results for many datasets. Additionally, they do not require advanced knowledge of model architectures, they take into account hyperparameter correlations and they work seamlessly across different hardware setups. All these reasons make them a strong option for the early stage of your experimentation process.
 
 ### Primary metric
 
 An AutoML training job uses a primary metric for model optimization and hyperparameter tuning. The primary metric depends on the task type as shown below; other primary metric values are currently not supported. 
 
-* Accuracy for image classification
-* Intersection over union for image classification multilabel
-* Mean average precision for image object detection
-* Mean average precision for image instance segmentation
+* [Accuracy](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.accuracy_score.html) for image classification
+* [Intersection over union]((https://scikit-learn.org/stable/modules/generated/sklearn.metrics.jaccard_score.html#sklearn.metrics.jaccard_score)) for image classification multilabel
+* [Mean average precision](en-us/azure/machine-learning/how-to-understand-automated-ml#object-detection-and-instance-segmentation-metrics) for image object detection
+* [Mean average precision](en-us/azure/machine-learning/how-to-understand-automated-ml#object-detection-and-instance-segmentation-metrics) for image instance segmentation
 
 ### Job limits
 
 You can control the resources spent on your AutoML Image training job by specifying the `timeout_minutes`, `max_trials` and the `max_concurrent_trials` for the job in limit settings as described in the below example.
 
 Parameter | Detail
 -----|----
-`max_trials` |  Parameter for maximum number of configurations to sweep. Must be an integer between 1 and 1000. When exploring just the default hyperparameters for a given model architecture, set this parameter to 1. The default value is 1.
-`max_concurrent_trials`| Maximum number of runs that can run concurrently. If specified, must be an integer between 1 and 100.  The default value is 1. <br><br> **NOTE:** <li> The number of concurrent runs is gated on the resources available in the specified compute target. Ensure that the compute target has the available resources for the desired concurrency.  <li> `max_concurrent_trials` is capped at `max_trials` internally. For example, if user sets `max_concurrent_trials=4`, `max_trials=2`, values would be internally updated as `max_concurrent_trials=2`, `max_trials=2`.
+`max_trials` |  Parameter for maximum number of trials to sweep. Must be an integer between 1 and 1000. When exploring just the default hyperparameters for a given model architecture, set this parameter to 1. The default value is 1.
+`max_concurrent_trials`| Maximum number of trials that can run concurrently. If specified, must be an integer between 1 and 100.  The default value is 1. <br><br> **NOTE:** <li> The number of concurrent trials is gated on the resources available in the specified compute target. Ensure that the compute target has the available resources for the desired concurrency.  <li> `max_concurrent_trials` is capped at `max_trials` internally. For example, if user sets `max_concurrent_trials=4`, `max_trials=2`, values would be internally updated as `max_concurrent_trials=2`, `max_trials=2`.
 `timeout_minutes`| The amount of time in minutes before the experiment terminates. If none specified, default experiment timeout_minutes is seven days (maximum 60 days)
 
 # [Azure CLI](#tab/cli)
@@ -324,7 +324,7 @@ limits:
 > [!IMPORTANT]
 > This feature is currently in public preview. This preview version is provided without a service-level agreement. Certain features might not be supported or might have constrained capabilities. For more information, see [Supplemental Terms of Use for Microsoft Azure Previews](https://azure.microsoft.com/support/legal/preview-supplemental-terms/).
 
-It is generally hard to predict the best model architecture and hyperparameters for a dataset. Also, in some cases the human time allocated to tuning hyperparameters may be limited. For computer vision tasks, you can specify a number of runs and the system will automatically determine the region of the hyperparameter space to sweep. You do not have to define a hyperparameter search space, a sampling method or an early termination policy.
+It is generally hard to predict the best model architecture and hyperparameters for a dataset. Also, in some cases the human time allocated to tuning hyperparameters may be limited. For computer vision tasks, you can specify a number of trials and the system will automatically determine the region of the hyperparameter space to sweep. You do not have to define a hyperparameter search space, a sampling method or an early termination policy.
 
 #### Triggering AutoMode
 
@@ -349,15 +349,15 @@ image_object_detection_job.set_limits(max_trials=10, max_concurrent_trials=2)
 ```
 ---
 
-A number of runs between 10 and 20 will likely work well on many datasets. The [time budget](#job-limits) for the AutoML job can still be set, but we recommend doing this only if each trial may take a long time.
+A number of trials between 10 and 20 will likely work well on many datasets. The [time budget](#job-limits) for the AutoML job can still be set, but we recommend doing this only if each trial may take a long time.
 
 > [!Warning]
 > Launching automatic sweeps via the UI is not supported at this time.
 
 
-### Individual runs
+### Individual trials
 
-In individual runs, you directly control the model architecture and hyperparameters. The model architecture is passed via the `model_name` parameter.
+In individual trials, you directly control the model architecture and hyperparameters. The model architecture is passed via the `model_name` parameter.
 
 #### Supported model architectures
 
@@ -441,7 +441,7 @@ search_space:
 
 You can define the model architectures and hyperparameters to sweep in the parameter space. You can either specify a single model architecture or multiple ones. 
 
-* See [Individual runs](#individual-runs) for the list of supported model architectures for each task type. 
+* See [Individual trials](#individual-trials) for the list of supported model architectures for each task type. 
 * See [Hyperparameters for computer vision tasks](reference-automl-images-hyperparameters.md)  hyperparameters for each computer vision task type. 
 * See [details on supported distributions for discrete and continuous hyperparameters](how-to-tune-hyperparameters.md#define-the-search-space).
 
@@ -460,7 +460,7 @@ When sweeping hyperparameters, you need to specify the sampling method to use fo
 
 #### Early termination policies
 
-You can automatically end poorly performing runs with an early termination policy. Early termination improves computational efficiency, saving compute resources that would have been otherwise spent on less promising configurations. Automated ML for images supports the following early termination policies using the `early_termination` parameter. If no termination policy is specified, all configurations are run to completion.
+You can automatically end poorly performing trials with an early termination policy. Early termination improves computational efficiency, saving compute resources that would have been otherwise spent on less promising trials. Automated ML for images supports the following early termination policies using the `early_termination` parameter. If no termination policy is specified, all trials are run to completion.
 
 
 | Early termination policy | AutoML Job syntax |
@@ -607,12 +607,12 @@ In our experiments, we found that these augmentations help the model to generali
 
 ##  Incremental training (optional)
 
-Once the training run is done, you have the option to further train the model by loading the trained model checkpoint. You can either use the same dataset or a different one for incremental training. 
+Once the training job is done, you have the option to further train the model by loading the trained model checkpoint. You can either use the same dataset or a different one for incremental training. 
 
 
-### Pass the checkpoint via run ID
+### Pass the checkpoint via job ID
 
-You can pass the run ID that you want to load the checkpoint from.
+You can pass the job ID that you want to load the checkpoint from.
 
 # [Azure CLI](#tab/cli)
 
@@ -628,10 +628,10 @@ training_parameters:
 
  [!INCLUDE [sdk v2](../../includes/machine-learning-sdk-v2.md)]
 
-To find the run ID from the desired model, you can use the following code. 
+To find the job ID from the desired model, you can use the following code. 
 
 ```python
-# find a run id to get a model checkpoint from
+# find a job id to get a model checkpoint from
 import mlflow
 
 # Obtain the tracking URL from MLClient
@@ -645,11 +645,11 @@ from mlflow.tracking.client import MlflowClient
 mlflow_client = MlflowClient()
 mlflow_parent_run = mlflow_client.get_run(automl_job.name)
 
-# Fetch the id of the best automl child run.
+# Fetch the id of the best automl child trial.
 target_checkpoint_run_id = mlflow_parent_run.data.tags["automl_best_child_run_id"]
 ```
 
-To pass a checkpoint via the run ID, you need to use the `checkpoint_run_id` parameter in `set_training_parameters` function.
+To pass a checkpoint via the job ID, you need to use the `checkpoint_run_id` parameter in `set_training_parameters` function.
 
 ```python
 image_object_detection_job = automl.image_object_detection(
@@ -697,18 +697,18 @@ When you've configured your AutoML Job to the desired settings, you can submit t
 
 ## Outputs and evaluation metrics
 
-The automated ML training runs generates output model files, evaluation metrics, logs and deployment artifacts like the scoring file and the environment file which can be viewed from the outputs and logs and metrics tab of the child runs.
+The automated ML training jobs generates output model files, evaluation metrics, logs and deployment artifacts like the scoring file and the environment file which can be viewed from the outputs and logs and metrics tab of the child jobs.
 
 > [!TIP]
-> Check how to navigate to the run results from the  [View run results](how-to-understand-automated-ml.md#view-job-results) section.
+> Check how to navigate to the job results from the  [View job results](how-to-understand-automated-ml.md#view-job-results) section.
 
-For definitions and examples of the performance charts and metrics provided for each run, see [Evaluate automated machine learning experiment results](how-to-understand-automated-ml.md#metrics-for-image-models-preview).
+For definitions and examples of the performance charts and metrics provided for each job, see [Evaluate automated machine learning experiment results](how-to-understand-automated-ml.md#metrics-for-image-models-preview).
 
 ## Register and deploy model
 
-Once the run completes, you can register the model that was created from the best run (configuration that resulted in the best primary metric). You can either register the model after downloading or by specifying the azureml path with corresponding jobid.  Note: If you want to change the inference settings that are described below you need to download the model and change settings.json and register using the updated model folder.
+Once the job completes, you can register the model that was created from the best trial (configuration that resulted in the best primary metric). You can either register the model after downloading or by specifying the azureml path with corresponding jobid.  Note: If you want to change the inference settings that are described below you need to download the model and change settings.json and register using the updated model folder.
 
-### Get the best run
+### Get the best trial
 
 # [Azure CLI](#tab/cli)
 
@@ -864,7 +864,7 @@ az ml online-endpoint update --name 'od-fridge-items-endpoint' --traffic 'od-fri
 
 
 Alternatively You can deploy the model from the [Azure Machine Learning studio UI](https://ml.azure.com/). 
-Navigate to the model you wish to deploy in the **Models** tab of the automated ML run and select on **Deploy** and select **Deploy to real-time endpoint** .  
+Navigate to the model you wish to deploy in the **Models** tab of the automated ML job and select on **Deploy** and select **Deploy to real-time endpoint** .  
 
 ![Screenshot of how the Deployment page looks like after selecting the Deploy option.](./media/how-to-auto-train-image-models/deploy-end-point.png).
 
@@ -1102,7 +1102,7 @@ image_object_detection_job = automl.image_object_detection(
 
 ### Streaming image files from storage
 
-By default, all image files are downloaded to disk prior to model training. If the size of the image files is greater than available disk space, the run will fail. Instead of downloading all images to disk, you can select to stream image files from Azure storage as they're needed during training. Image files are streamed from Azure storage directly to system memory, bypassing disk. At the same time, as many files as possible from storage are cached on disk to minimize the number of requests to storage.
+By default, all image files are downloaded to disk prior to model training. If the size of the image files is greater than available disk space, the job will fail. Instead of downloading all images to disk, you can select to stream image files from Azure storage as they're needed during training. Image files are streamed from Azure storage directly to system memory, bypassing disk. At the same time, as many files as possible from storage are cached on disk to minimize the number of requests to storage.
 
 > [!NOTE]
 > If streaming is enabled, ensure the Azure storage account is located in the same region as compute to minimize cost and latency.

articles/machine-learning/how-to-inference-onnx-automl-image-models.md

@@ -1144,7 +1144,7 @@ for image_idx, class_idx in zip(image_wise_preds[0], image_wise_preds[1]):
     print('image: {}, class_index: {}, class_name: {}'.format(image_files[image_idx], class_idx, classes[class_idx]))
 ```
 
-For multi-class and multi-label classification, you can follow the same steps mentioned earlier for all the supported architectures in AutoML.
+For multi-class and multi-label classification, you can follow the same steps mentioned earlier for all the supported model architectures in AutoML.
 
 
 # [Object detection with Faster R-CNN or RetinaNet](#tab/object-detect-cnn)