The parameters present in a config.json file allow one to configure a session. Each of these parameters is described below:
Where to find and how to handle the data.
- path: string Path to dataset
- factor: int Value of the factor applied during downsampling and upsampling
- xSamples: int Number of ground truth samples along the x-axis in the room. (32 if using the provided dataset)
- ySamples: int Number of ground truth samples along the y-axis in the room. (32 if using the provided dataset)
How training will be performed.
- batch_size: int Number of samples in a batch
- num_epochs: int Number of epochs to train for
- num_steps_val: int Total number of steps (batches of samples) to yield from validation generator before stopping at the end of every epoch.
- num_steps_train: int Total number of steps (batches of samples) to yield from validation generator before stopping at the end of every epoch.
- session_id: int Numerical identifier for this session
- lr: float Learning rate
- loss: float
- valid_weight: float Weight given to the loss term considering microphone position predictions.
- hole_weight: float Weight given to the loss term considering non-microphone position predictions.
How evaluation will be performed.
- min_mics: int Minimum number of microphones placed in a room to evaluate the model.
- max_mics: int Maximum+1 number of microphones placed in a room to evaluate the model.
- step_mics: int Spacing between the value of the number of microphones placed.
- num_comb: int Number of different irregular patterns tested with a fix amount of microphones.
How visualization will be performed.
- num_mics: int Number of microphones randomly located in the real room.
- source: int Numerical identifier of the source location. Must be either 0 or 1.