retinanetjs

load

• load(modelPath: string | IOHandler, classes: string[], preprocessingMode: string, onProgress?: function, anchorParams?: AnchorParameters): Promise<RetinaNet>

• Parameters

  • modelPath: string | IOHandler

    The path to the model or a tf.io.IOHandler object

  • classes: string[]

    The list of detected classes

  • preprocessingMode: string

    One of tf or caffe. Check the preprocess_images method of your backbone to see which you should use.

  • Optional onProgress: function

    A callback to report progress

    • • (progress: number, message: string): void

      • Parameters

        • progress: number

        • message: string

        Returns void

  • Default value anchorParams: AnchorParameters = defaultAnchorParameters

    The anchor parameters for your model

  Returns Promise<RetinaNet>