name: "BNR2" model: type: bnr sample_rate: 16000 fft_length: 1920 hop_length: 480 num_mels: 320 skip_nan_grad: false num_outputs: 1 segment: 4 train: # Parameters related to training enable_weight_norm: true optim: name: adam lr: 0.0005 sched: name: StepLR gamma: 0.999 step_size: 2 train_ds: manifest_filepath: ??? input_key: noisy_filepath # key of the input signal path in the manifest target_key: speech_filepath # key of the target signal path in the manifest target_channel_selector: 0 # target signal is the first channel from files in target_key audio_duration: 4.0 # in seconds, audio segment duration for training random_offset: true # if the file is longer than audio_duration, use random offset to select a subsegment min_duration: ${model.train_ds.audio_duration} batch_size: 64 # batch size may be increased based on the available memory shuffle: true num_workers: 8 pin_memory: true validation_ds: manifest_filepath: ??? input_key: noisy_filepath # key of the input signal path in the manifest target_key: speech_filepath # key of the target signal path in the manifest target_channel_selector: 0 # target signal is the first channel from files in target_key audio_duration: 10.0 # in seconds, audio segment duration for validation min_duration: ${model.validation_ds.audio_duration} batch_size: 64 # batch size may be increased based on the available memory shuffle: false num_workers: 4 pin_memory: true test_ds: manifest_filepath: ??? input_key: noisy_filepath # key of the input signal path in the manifest target_key: speech_filepath # key of the target signal path in the manifest target_channel_selector: 0 # target signal is the first channel from files in target_key audio_duration: 10.0 # in seconds, audio segment duration for validation min_duration: ${model.test_ds.audio_duration} batch_size: 1 # batch size may be increased based on the available memory shuffle: false num_workers: 4 pin_memory: true loss: _target_: nemo.collections.audio.losses.maxine.CombinedLoss sample_rate: ${model.sample_rate} fft_length: ${model.fft_length} hop_length: ${model.hop_length} num_mels: ${model.num_mels} sisnr_loss_weight: 1 spectral_loss_weight: 15 asr_loss_weight: 1 use_asr_loss: true use_mel_spec: true metrics: val: sdr: # output SDR _target_: torchmetrics.audio.SignalDistortionRatio test: sdr_ch0: # SDR on output channel 0 _target_: torchmetrics.audio.SignalDistortionRatio channel: 0 trainer: devices: -1 # number of GPUs, -1 would use all available GPUs num_nodes: 1 max_epochs: -1 max_steps: -1 # computed at runtime if not set val_check_interval: 1.0 # Set to 0.25 to check 4 times per epoch, or an int for number of iterations accelerator: gpu strategy: ddp accumulate_grad_batches: 1 gradient_clip_val: 5 precision: 32 # Should be set to 16 for O1 and O2 to enable the AMP. log_every_n_steps: 25 # Interval of logging. enable_progress_bar: true num_sanity_val_steps: 0 # number of steps to perform validation steps for sanity check the validation process before starting the training, setting to 0 disables it check_val_every_n_epoch: 1 # number of evaluations on validation every n epochs sync_batchnorm: true enable_checkpointing: False # Provided by exp_manager logger: false # Provided by exp_manager exp_manager: exp_dir: null name: ${name} create_tensorboard_logger: true create_checkpoint_callback: true checkpoint_callback_params: # in case of multiple validation sets, first one is used monitor: "val_loss" mode: "min" save_top_k: 5 always_save_nemo: true # saves the checkpoints as nemo files instead of PTL checkpoints resume_from_checkpoint: null # The path to a checkpoint file to continue the training, restores the whole state including the epoch, step, LR schedulers, apex, etc. # you need to set these two to true to continue the training resume_if_exists: false resume_ignore_no_checkpoint: false # You may use this section to create a W&B logger create_wandb_logger: false wandb_logger_kwargs: name: null project: null