You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
PaddleSpeech/examples/audio/codec/dac/conf/default.yaml

38 lines
801 B

# Configuration for DAC model
# Model Configuration
model:
sample_rate: 44100
encoder_dims: 512
decoder_dims: 512
num_residual_layers: 10
n_fft: 1024
hop_length: 256
# Training Configuration
training:
batch_size: 16
lr: 0.0001
weight_decay: 0.0001
gradient_clip_val: 1.0
max_epochs: 200
warmup_steps: 1000
# Loss weights
recon_loss_weight: 1.0
mel_loss_weight: 10.0
adversarial_loss_weight: 0.1
# Data Configuration
data:
train_manifest: "data/train/manifest.json"
dev_manifest: "data/dev/manifest.json"
test_manifest: "data/test/manifest.json"
max_duration: 10.0 # Maximum audio length in seconds
min_duration: 1.0 # Minimum audio length in seconds
# Distributed Training Configuration
distributed:
world_size: 1
find_unused_parameters: true