definition.yaml

author_notes: null
benchmark:
  benchmark_description: please note that fluent-speech-corpus dataset hosted on Kaggle
    is a licensed dataset.
  benchmark_link: https://www.kaggle.com/tommyngx/fluent-speech-corpus
  benchmark_metrics:
    LER: '0.0348'
    WER: '0.1123'
  benchmark_name: Fluent speech
description: "Tiny Wav2letter is a tiny version of the original Wav2Letter model.\
  \ It is a convolutional speech recognition neural network. This implementation was\
  \ created by Arm, pruned to 50% sparsity, fine-tuned and quantized using the TensorFlow\
  \ Model Optimization Toolkit.\r\n\r\n"
license:
- Apache-2.0
network:
  datatype: int8
  file_size_bytes: 3997112
  filename: tiny_wav2letter_int8.tflite
  framework: TensorFlow Lite
  framework_version: 2.4.1
  hash:
    algorithm: sha1
    value: 13ca2294ba4bbb1f1c6c5e663cb532d58cd76a6b
  provenance: https://github.com/ARM-software/ML-zoo/tree/master/models/speech_recognition/wav2letter
  training: LibriSpeech,Mini LibrySpeech,fluent speech
network_parameters:
  input_nodes:
  - description: Speech converted to MFCCs and quantized to INT8
    example_input:
      path: models/speech_recognition/tiny_wav2letter/tflite_int8/testing_input/input_1_int8
    input_datatype: int8
    name: input_1_int8
    shape:
    - 1
    - 296
    - 39
  output_nodes:
  - description: A tensor of time and class probabilities, that represents the probability
      of each class at each timestep. Should be passed to a decoder. For example ctc_beam_search_decoder.
    example_output:
      path: models/speech_recognition/tiny_wav2letter/tflite_int8/testing_output/Identity_int8
    name: Identity_int8
    output_datatype: int8
    shape:
    - 1
    - 1
    - 148
    - 29
network_quality:
  quality_level: Deployable
  quality_level_hero_hw: null
operators:
  TensorFlow Lite:
  - CONV_2D
  - DEQUANTIZE
  - LEAKY_RELU
  - QUANTIZE
  - RESHAPE
paper: https://arxiv.org/abs/1609.03193