definition.yaml

benchmark:
  Visual Wake Words:
    accuracy: 0.768
description: 'This is a fully quantized version (asymmetrical int8) of the MicroNet
  VWW-2 model developed by Arm, from the MicroNets paper. It is trained on the ''Visual
  Wake Words'' dataset, more information can be found here: https://arxiv.org/pdf/1906.05721.pdf.'
license:
- Apache-2.0
network:
  file_size_bytes: 280384
  filename: vww2_50_50_INT8.tflite
  framework: TensorFlow Lite
  hash:
    algorithm: sha1
    value: 5d887ca438c0a7feeed3c8c22dce99b55565c8ea
  provenance: https://arxiv.org/pdf/2010.11267.pdf
network_parameters:
  input_nodes:
  - description: A 50x50 input image.
    example_input:
      path: models/visual_wake_words/micronet_vww2/tflite_int8/testing_input/input
    name: input
    shape:
    - 1
    - 50
    - 50
    - 1
  output_nodes:
  - description: Per-class confidence across the two classes (0=no person present,
      1=person present).
    name: Identity
    shape:
    - 1
    - 2
    test_output_path: models/visual_wake_words/micronet_vww2/tflite_int8/testing_output/Identity
operators:
  TensorFlow Lite:
  - ADD
  - AVERAGE_POOL_2D
  - CONV_2D
  - DEPTHWISE_CONV_2D
  - PAD
  - RELU6
  - RESHAPE
paper: https://arxiv.org/pdf/2010.11267.pdf