ComVo / config.yaml
hsoh's picture
Upload 2 files
9bbd232 verified
# pytorch_lightning==2.5.0
seed_everything: 4444
trainer:
accelerator: gpu
strategy: auto
devices: -1
num_nodes: 1
precision: null
logger:
class_path: pytorch_lightning.loggers.TensorBoardLogger
init_args:
save_dir: ./ComVo
name: lightning_logs
version: null
log_graph: false
default_hp_metric: true
prefix: ''
sub_dir: null
comment: ''
purge_step: null
max_queue: 10
flush_secs: 120
filename_suffix: ''
callbacks:
- class_path: pytorch_lightning.callbacks.LearningRateMonitor
init_args:
logging_interval: null
log_momentum: false
log_weight_decay: false
- class_path: pytorch_lightning.callbacks.ModelSummary
init_args:
max_depth: 2
- class_path: pytorch_lightning.callbacks.ModelCheckpoint
init_args:
dirpath: null
filename: checkpoint_{epoch}_{step}_{val_loss:.4f}
monitor: val_loss
verbose: false
save_last: true
save_top_k: 3
save_weights_only: false
mode: min
auto_insert_metric_name: true
every_n_train_steps: null
train_time_interval: null
every_n_epochs: null
save_on_train_epoch_end: null
enable_version_counter: true
- class_path: exp.helpers.GradNormCallback
fast_dev_run: false
max_epochs: null
min_epochs: null
max_steps: 2000000
min_steps: null
max_time: null
limit_train_batches: null
limit_val_batches: 100
limit_test_batches: null
limit_predict_batches: null
overfit_batches: 0.0
val_check_interval: null
check_val_every_n_epoch: 1
num_sanity_val_steps: null
log_every_n_steps: 100
enable_checkpointing: null
enable_progress_bar: null
enable_model_summary: null
accumulate_grad_batches: 1
gradient_clip_val: null
gradient_clip_algorithm: null
deterministic: null
benchmark: null
inference_mode: true
use_distributed_sampler: true
profiler: null
detect_anomaly: false
barebones: false
plugins: null
sync_batchnorm: false
reload_dataloaders_every_n_epochs: 0
default_root_dir: null
model:
class_path: exp.experiment_cdisc.ComVoExp_cdisc
init_args:
feature_extractor:
class_path: exp.feature_extractors.MelSpectrogramFeatures
init_args:
sample_rate: 24000
n_fft: 1024
hop_length: 256
n_mels: 100
padding: center
backbone:
class_path: exp.models.ComVo
init_args:
input_channels: 100
dim: 512
intermediate_dim: 1536
num_layers: 8
n_quantization: 128
layer_scale_init_value: null
adanorm_num_embeddings: null
rank: null
head:
class_path: exp.heads.ISTFTHead
init_args:
dim: 512
n_fft: 1024
hop_length: 256
padding: center
sample_rate: 24000
initial_learning_rate: 0.0002
num_warmup_steps: 0
mel_loss_coeff: 45.0
mrd_loss_coeff: 0.1
pretrain_mel_steps: 0
decay_mel_coeff: false
evaluate_utmos: true
evaluate_pesq: true
evaluate_periodicty: true
data:
class_path: exp.dataset.AudioDataModule
init_args:
train_params:
filelist_path: filelist.train
sampling_rate: 24000
num_samples: 16384
batch_size: 16
num_workers: 8
val_params:
filelist_path: filelist.val
sampling_rate: 24000
num_samples: 48384
batch_size: 16
num_workers: 8
ckpt_path: null
optimizer: null
lr_scheduler: null