# AudioTranscriberModelArchitecture

    enum AudioTranscriberModelArchitecture: RawRepresentable, Sendable, Equatable, Hashable, BitwiseCopyable

Model architecture backend family for `MPPAudioTranscriber`.
- `
  ### [auto](https://developers.google.com/edge/api/mediapipe/swift/audio/Enums/AudioTranscriberModelArchitecture.md.txt#/c:_E_MPPAudioTranscriberModelArchitecture_MPPAudioTranscriberModelArchitectureAuto)

  ` Automatically infer the model architecture from the model asset.

  #### Declaration

  Swift

      case auto

- `
  ### [omni](https://developers.google.com/edge/api/mediapipe/swift/audio/Enums/AudioTranscriberModelArchitecture.md.txt#/c:_E_MPPAudioTranscriberModelArchitecture_MPPAudioTranscriberModelArchitectureOmni)

  ` Open-weight ASR models executed via the LiteRT Omni ASR runtime.

  #### Declaration

  Swift

      case omni

- `
  ### [tinyGemma](https://developers.google.com/edge/api/mediapipe/swift/audio/Enums/AudioTranscriberModelArchitecture.md.txt#/c:_E_MPPAudioTranscriberModelArchitecture_MPPAudioTranscriberModelArchitectureTinyGemma)

  ` First-party Tiny Gemma ASR models executed via the LiteRT-LM engine.

  #### Declaration

  Swift

      case tinyGemma

- `
  ### [init(rawValue:)](https://developers.google.com/edge/api/mediapipe/swift/audio/Enums/AudioTranscriberModelArchitecture.md.txt#/s:So36MPPAudioTranscriberModelArchitectureV8rawValueABSgSu_tcfc)

  `

  #### Declaration

  Swift

      init?(rawValue: UInt)