Train a custom model with your dataset using SmartML
Start training computer vision models right from Plainsight using your labeled datasets. With our SmartML model training backend, you can easily begin a training run with a click of your mouse, no coding required.
Starting a Training Run
To begin, you will need a locked version of your dataset. See Versioning for more information.
After you've created and locked a dataset version, you'll be able to train a model using your labeled data.
Go to "Models" in the side nav
Click "Add Model" at the top right, or select an existing model and click "New Version"
If this is a new model, enter a name for your model.
Select the desired model options for your training run:
Labels to be used in training. A model is trained on only one label type.
Only Bounding box (rectangle labels) and Instance Segmentation are currently supported in SmartML training. Other label types (Feature Points, Text, Classification, and MultiClassification) can be exported and trained outside of Plainsight.
Training length, between 1-24 hrs. This is the amount of time spent training your model. Read more about Training Time.
Your model will automatically stop training when it stops measurably improving or when the training reaches the budgeted time.
Your model version will be created and displayed at the top of your model's version list.
When a model enters the training process, it goes through several states:
Starting - Training job is waiting to begin
Running - Training is actively running
Succeeded - Training is complete
Failed - Training encountered an error and was stopped prematurely by the system
Canceled - Training run was stopped prematurely by the user
Stopping a Training Run
You can stop a current training run from the Model Version Details screen of the actively training model:
or from the Model Versions list:
The model training time is constrained to the selected training length that is specified in the model options. Actual time used may be a few minutes higher as there is some setup that occurs before the training process starts.
When a model finishes training, the Training Time is listed in 2 parts:
Used time - The total amount of time the training process takes from beginning to end, including pre and post-processing (initializing a container). This time involves container setup and generating predictions on the dataset images. Unpredictable processing times may cause slight overages especially for larger datasets (10-15%).
Budgeted Time - this is the length that was specified in your model training options. SmartML will automatically stop training when your model's validation loss stops decreasing, saving unnecessary training hours.
Successfully trained models can then be deployed for inferencing and testing.
Dataset Image Viewer
After training your model, you can view the detections on your dataset images and compare it to the labeled annotations. At the bottom of the model version details, you can preview the splits in Test Set, Validation Set, and Train Set.
Select the tab for test Set, Validation Set, or Train Set.
Click on an image thumbnail to open the image.
Under Display, select "Model Detection" to see the labels that the model detected. Select "Labeled Annotations" to see what was actually labeled.