SetClassifierTrainingOption

Sets the value of a training option that is applied to a classifier.

You can set the following training options:

Key Description Default
batchsize The number of images to train concurrently. The value must be 1,2,4,8,16, or 32. Larger batch sizes result in slightly faster training but require more memory (more GPU memory if you are training with a GPU). If you are training with a GPU that has limited memory you might need to reduce the batch size. 32
iterations The number of training iterations to perform. A larger number of iterations can result in more accurate classification, especially for classifiers that contain many similar classes. 500
snapshot_frequency The number of training iterations to perform before taking a snapshot of the classifier. For example, if you run 1000 iterations you might take snapshots every 250 iterations, so that there are snapshots for 250, 500, 750, and 1000 iterations. Taking snapshots can help you find the optimum number of iterations. No snapshots
validation_proportion

The proportion of your training images to set aside to evaluate the performance of the classifier. Specify a number between 0 and 1.

  • By default, Media Server does not evaluate the performance of a classifier and all data is used for training.
  • If you set validation_proportion but not snapshot_frequency, Media Server uses the specified proportion of your training images to evaluate the final trained model.
  • If you set snapshot_frequency, Media Server uses the specified proportion of your training images to evaluate all snapshots. The default validation_proportion uses 75% of your images for training and 25% for evaluation.
Zero, unless you set snapshot_frequency in which case 0.25.

NOTE: Changing a training option invalidates all training associated with a classifier. After using this action, you must retrain Media Server by running the action BuildClassifier.

Type: synchronous

Parameter Description Required
classifier The name of the classifier to configure. Yes
key The name of the training option to set. Yes
value The value for the training option. Yes

Example

This example sets the iterations training option to 250:

/action=SetClassifierTrainingOption&classifier=vehicles
                                   &key=iterations
                                   &value=250

Response

<autnresponse>
   <action>SETCLASSIFIERTRAININGOPTION</action>
   <response>SUCCESS</response>
   <responsedata></responsedata>
</autnresponse>