Refining models

After deploying a model, you can improve its accuracy by supplying more data. You can use several methods to add more data to the model.

You can add more data by using any combination of the following methods:

• Upload new images or videos to the data set and classify or label them as appropriate.
• For an existing video, capture more frames and classify or label them as appropriate. Or, for action detection models, label more actions.
• Use data augmentation. Data augmentation is the use of filters, such as blur and rotate, to create new versions of existing images or frames. Augmentation does not apply to full videos. It can be applied to a video’s captured frames in the same way that it is applied to images. When you use data augmentation, a new data set is created that contains all of the existing images plus the newly generated images, which are marked as augmented. For instructions, see Augmenting the data set.
• For models trained for object detection, use the Auto label function to identify more objects in the existing data. See Automatically labeling objects for instructions.
• When deploying an object detection model, choose Advanced deployment and specify that inference results are saved to a data set. Objects that are labeled this way have the type "inferred", and the label is green. You can accept or reject the inferred labels. Accepted labels are considered manually added and are changed to blue.After adding more data, train the model again.