turian/sgmse-speech-enhancement-deverb-replicate

turian / sgmse-speech-enhancement-deverb-replicate

Score-based Generative Models (Diffusion Models) for Speech Enhancement and Dereverberation

Public
59 runs
GitHub
Paper
License

Run turian/sgmse-speech-enhancement-deverb-replicate with an API

Use one of our client libraries to get started quickly. Clicking on a library will take you to the Playground tab where you can tweak different inputs, see the results, and copy the corresponding code to use in your own project.

Input schema

The fields you can use to run this model with an API. If you don't give a value for a field its default value will be used.

Field	Type	Default value	Description
audio	string		Speech audio file
checkpoint	string (enum)	EARS-WHAM Options: EARS-WHAM, EARS-Reverb	Model checkpoint to use. EARS-WHAM speech enhancement or EARS-Reverb dereverberation.
corrector	string (enum)	ald Options: ald, langevin, none	Corrector class for the PC sampler.
corrector_steps	integer	1	Number of corrector steps
snr	number	0.5	SNR value for (annealed) Langevin dynamics.
N	integer	30	Number of reverse steps

{
  "type": "object",
  "title": "Input",
  "required": [
    "audio"
  ],
  "properties": {
    "N": {
      "type": "integer",
      "title": "N",
      "default": 30,
      "x-order": 5,
      "description": "Number of reverse steps"
    },
    "snr": {
      "type": "number",
      "title": "Snr",
      "default": 0.5,
      "x-order": 4,
      "description": "SNR value for (annealed) Langevin dynamics."
    },
    "audio": {
      "type": "string",
      "title": "Audio",
      "format": "uri",
      "x-order": 0,
      "description": "Speech audio file"
    },
    "corrector": {
      "enum": [
        "ald",
        "langevin",
        "none"
      ],
      "type": "string",
      "title": "corrector",
      "description": "Corrector class for the PC sampler.",
      "default": "ald",
      "x-order": 2
    },
    "checkpoint": {
      "enum": [
        "EARS-WHAM",
        "EARS-Reverb"
      ],
      "type": "string",
      "title": "checkpoint",
      "description": "Model checkpoint to use. EARS-WHAM speech enhancement or EARS-Reverb dereverberation.",
      "default": "EARS-WHAM",
      "x-order": 1
    },
    "corrector_steps": {
      "type": "integer",
      "title": "Corrector Steps",
      "default": 1,
      "x-order": 3,
      "description": "Number of corrector steps"
    }
  }
}

Output schema

The shape of the response you’ll get when you run this model with an API.

Schema

{
  "type": "string",
  "title": "Output",
  "format": "uri"
}