Andberg9's workspace
Runs
36
Name
36 visualized
State
Notes
User
Tags
Created
Runtime
Sweep
adapter_type
augmentations.augmentations
augmentations.augmentations_per_sample
batch_size
betas
general_config
gradient_clipping_enabled
learning_rate
method
metric.goal
metric.name
name
num_classes
optimizer_config
parameters.adapter_types
parameters.batch_sizes
parameters.betass
parameters.gradient_clipping_enableds
parameters.learning_rates
parameters.peft_scheduling.schedule.0.peft_methods
parameters.peft_scheduling.schedule.0.start_epochs
parameters.peft_scheduling.schedule.1.peft_methods
parameters.peft_scheduling.schedule.1.start_epochs
parameters.weight_decays
peft_config
peft_scheduling.schedule.0.peft_method
peft_scheduling.schedule.0.start_epoch
peft_scheduling.schedule.1.peft_method
peft_scheduling.schedule.1.start_epoch
peft_scheduling_config
project
weight_decay
best_val_accuracy
epoch
inference_acc
inference_accuracy
inference_f1
inference_precision
inference_recall
lr-AdamW
model/memory_footprint_mb
model/total_parameters
model/trainable_parameters
model/trainable_percentage
Crashed
andberg9
15m 33s
-
["time_mask","time_stretch","sin_distortion"]
2
-
-
model_type='efficientnet_b0' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=2 pinned_memory=True epochs=30 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=200 accumulation_steps=1 patience=20 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='lorac' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=1 strategy='ddp'
-
-
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0008, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.01, amsgrad=False) warmup=WarmupConfig(enabled=True, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=1.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
adapter_type='lorac' task_type='SEQ_CLS' r=4 alpha=8.0 dropout=0.0 target_modules=['linear', 'conv2d', 'batchnorm2d']
none-classifier
0
none-full
5
enabled=True model_name=None schedule=[PEFTScheduleStep(start_epoch=0, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=5, peft_method='lora', merge_previous=True)] auto_merge=True
-
-
-
4
-
-
-
-
-
0.00004894
21.72163
5694194
1623172
28.50574
Crashed
andberg9
3m 47s
-
["time_mask","time_stretch","sin_distortion"]
0
-
-
model_type='efficientnet_b0' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=2 pinned_memory=True epochs=30 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=200 accumulation_steps=1 patience=20 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='lorac' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=1 strategy='ddp'
-
-
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0008, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.01, amsgrad=False) warmup=WarmupConfig(enabled=True, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=1.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
adapter_type='lorac' task_type='SEQ_CLS' r=4 alpha=8.0 dropout=0.0 target_modules=['linear', 'conv2d', 'batchnorm2d']
none-classifier
0
none-full
5
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=0, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=5, peft_method='lora', merge_previous=True)] auto_merge=True
-
-
-
1
-
-
-
-
-
0.00002497
21.72163
5694194
1623172
28.50574
Failed
andberg9
3s
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
none-classifier
0
none-full
5
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Failed
andberg9
2s
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
none-classifier
0
none-full
5
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Failed
andberg9
5s
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
none-classifier
0
none-full
5
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Failed
andberg9
2s
hra
-
-
8
[0.9,0.999]
-
false
0.0001
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
0
-
-
-
-
-
-
-
-
-
-
-
-
Failed
andberg9
1s
oft
-
-
8
[0.9,0.999]
-
false
0.0001
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
0.1
-
-
-
-
-
-
-
-
-
-
-
-
Crashed
andberg9
25m 2s
oft
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.99,0.999]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='oft' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
true
0.0005
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0005, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
OFTConfig(peft_type=<PeftType.OFT: 'OFT'>, auto_mapping=None, base_model_name_or_path='MIT/ast-finetuned-audioset-10-10-0.4593', revision=None, task_type='SEQ_CLS', inference_mode=False, rank_pattern={}, alpha_pattern={}, r=128, module_dropout=0.2, target_modules=['audio_spectrogram_transformer.encoder.layer.0.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.0.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.0.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.0.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.1.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.1.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.2.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.2.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.3.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.3.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.4.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.4.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.5.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.5.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.6.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.6.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.7.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.7.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.8.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.8.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.9.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.9.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.10.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.10.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.11.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.11.output.dense', 'classifier.dense'], init_weights=True, layers_to_transform=None, layers_pattern=None, modules_to_save=['classifier', 'score'], coft=False, eps=6e-05, block_share=False)
-
-
-
-
-
-
0
-
4
-
-
-
-
-
0.00048244
333.51405
87428708
1201330
1.37407
Finished
andberg9
1h 7m 29s
hra
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.9,0.999]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='hra' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
false
0.0001
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.1, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=False gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
HRAConfig(peft_type=<PeftType.HRA: 'HRA'>, auto_mapping=None, base_model_name_or_path='MIT/ast-finetuned-audioset-10-10-0.4593', revision=None, task_type='SEQ_CLS', inference_mode=False, r=2, apply_GS=False, target_modules=['audio_spectrogram_transformer.encoder.layer.0.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.0.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.0.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.0.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.1.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.1.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.2.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.2.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.3.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.3.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.4.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.4.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.5.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.5.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.6.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.6.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.7.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.7.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.8.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.8.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.9.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.9.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.10.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.10.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.11.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.11.output.dense', 'classifier.dense'], init_weights=True, layers_to_transform=None, layers_pattern=None, bias='none', modules_to_save=['classifier', 'score'])
-
-
-
-
-
-
0.1
0
20
0.135
0.135
0.11475
0.12781
0.135
0.000065451
329.72792
86436196
207410
0.23996
Finished
andberg9
1h 34m 31s
oft
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.9,0.999]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='oft' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
true
0.0001
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
OFTConfig(peft_type=<PeftType.OFT: 'OFT'>, auto_mapping=None, base_model_name_or_path='MIT/ast-finetuned-audioset-10-10-0.4593', revision=None, task_type='SEQ_CLS', inference_mode=False, rank_pattern={}, alpha_pattern={}, r=128, module_dropout=0.2, target_modules=['audio_spectrogram_transformer.encoder.layer.0.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.0.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.0.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.0.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.1.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.1.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.2.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.2.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.3.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.3.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.4.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.4.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.5.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.5.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.6.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.6.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.7.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.7.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.8.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.8.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.9.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.9.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.10.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.10.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.11.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.11.output.dense', 'classifier.dense'], init_weights=True, layers_to_transform=None, layers_pattern=None, modules_to_save=['classifier', 'score'], coft=False, eps=6e-05, block_share=False)
-
-
-
-
-
-
0
0
20
0.7775
0.7775
0.76419
0.77674
0.7775
0.000065451
333.51405
87428708
1201330
1.37407
Finished
andberg9
39m 35s
layernorm
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.9,0.999]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='layernorm' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
true
0.0005
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0005, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.1, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
LNTuningConfig(peft_type=<PeftType.LN_TUNING: 'LN_TUNING'>, auto_mapping=None, base_model_name_or_path='MIT/ast-finetuned-audioset-10-10-0.4593', revision=None, task_type='SEQ_CLS', inference_mode=False, target_modules=['audio_spectrogram_transformer.encoder.layer.0.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.0.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.0.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.0.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.1.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.1.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.2.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.2.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.3.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.3.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.4.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.4.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.5.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.5.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.6.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.6.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.7.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.7.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.8.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.8.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.9.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.9.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.10.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.10.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.11.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.11.output.dense', 'classifier.dense'], modules_to_save=['classifier', 'score'])
-
-
-
-
-
-
0.1
0
18
0.4375
0.4375
0.42886
0.47539
0.4375
0.00035644
653.69315
171361736
85096036
49.65871
Finished
andberg9
43m 52s
none-full
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.99,0.999]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
true
0.0005
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0005, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.1, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
adapter_type='none-full' task_type='SEQ_CLS'
-
-
-
-
-
-
0.1
0
20
0.4725
0.4725
0.46497
0.51397
0.4725
0.00032725
328.93085
86227250
86227250
100
Finished
andberg9
17m 22s
none-classifier
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.85,0.89]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-classifier' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
true
0.0005
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0005, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
adapter_type='none-classifier' task_type='SEQ_CLS'
-
-
-
-
-
-
0
0
16
0.6125
0.6125
0.60393
0.63627
0.6125
0.00038396
328.93085
86227250
39986
0.046373
Finished
andberg9
1h 7m 28s
hra
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.9,0.999]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='hra' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
true
0.0001
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.01, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
HRAConfig(peft_type=<PeftType.HRA: 'HRA'>, auto_mapping=None, base_model_name_or_path='MIT/ast-finetuned-audioset-10-10-0.4593', revision=None, task_type='SEQ_CLS', inference_mode=False, r=2, apply_GS=False, target_modules=['audio_spectrogram_transformer.encoder.layer.0.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.0.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.0.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.0.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.1.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.1.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.2.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.2.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.3.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.3.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.4.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.4.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.5.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.5.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.6.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.6.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.7.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.7.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.8.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.8.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.9.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.9.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.10.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.10.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.11.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.11.output.dense', 'classifier.dense'], init_weights=True, layers_to_transform=None, layers_pattern=None, bias='none', modules_to_save=['classifier', 'score'])
-
-
-
-
-
-
0.01
0
20
0.125
0.125
0.11049
0.12618
0.125
0.000065451
329.72792
86436196
207410
0.23996
Finished
andberg9
47m 58s
oft
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.85,0.89]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='oft' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
false
0.0005
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0005, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.01, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=False gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
OFTConfig(peft_type=<PeftType.OFT: 'OFT'>, auto_mapping=None, base_model_name_or_path='MIT/ast-finetuned-audioset-10-10-0.4593', revision=None, task_type='SEQ_CLS', inference_mode=False, rank_pattern={}, alpha_pattern={}, r=128, module_dropout=0.2, target_modules=['audio_spectrogram_transformer.encoder.layer.0.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.0.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.0.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.0.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.1.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.1.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.2.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.2.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.3.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.3.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.4.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.4.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.5.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.5.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.6.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.6.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.7.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.7.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.8.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.8.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.9.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.9.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.10.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.10.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.11.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.11.output.dense', 'classifier.dense'], init_weights=True, layers_to_transform=None, layers_pattern=None, modules_to_save=['classifier', 'score'], coft=False, eps=6e-05, block_share=False)
-
-
-
-
-
-
0.01
0
10
0.785
0.785
0.77337
0.78324
0.785
0.00045225
333.51405
87428708
1201330
1.37407
Finished
andberg9
34m 57s
layernorm
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.8,0.85]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='layernorm' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
true
0.0001
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.01, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
LNTuningConfig(peft_type=<PeftType.LN_TUNING: 'LN_TUNING'>, auto_mapping=None, base_model_name_or_path='MIT/ast-finetuned-audioset-10-10-0.4593', revision=None, task_type='SEQ_CLS', inference_mode=False, target_modules=['audio_spectrogram_transformer.encoder.layer.0.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.0.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.0.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.0.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.1.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.1.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.2.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.2.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.3.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.3.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.4.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.4.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.5.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.5.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.6.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.6.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.7.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.7.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.8.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.8.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.9.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.9.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.10.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.10.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.11.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.11.output.dense', 'classifier.dense'], modules_to_save=['classifier', 'score'])
-
-
-
-
-
-
0.01
0
16
0.7225
0.7225
0.71828
0.73708
0.7225
0.000076791
653.69315
171361736
85096036
49.65871
Finished
andberg9
35m 8s
lora
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.85,0.89]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='lora' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
false
0.0001
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.01, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=False gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
LoraConfig(peft_type=<PeftType.LORA: 'LORA'>, auto_mapping=None, base_model_name_or_path='MIT/ast-finetuned-audioset-10-10-0.4593', revision=None, task_type='SEQ_CLS', inference_mode=False, r=8, target_modules=['audio_spectrogram_transformer.encoder.layer.0.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.0.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.0.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.0.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.1.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.1.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.2.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.2.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.3.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.3.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.4.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.4.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.5.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.5.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.6.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.6.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.7.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.7.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.8.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.8.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.9.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.9.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.10.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.10.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.11.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.11.output.dense', 'classifier.dense'], lora_alpha=16, lora_dropout=0, fan_in_fan_out=False, bias='lora_only', use_rslora=False, modules_to_save=['classifier', 'score'], init_lora_weights=True, layers_to_transform=None, layers_pattern=None, rank_pattern={}, alpha_pattern={}, megatron_config=None, megatron_core='megatron.core', loftq_config={}, use_dora=False, layer_replication=None, runtime_config=LoraRuntimeConfig(ephemeral_gpu_offload=False))
-
-
-
-
-
-
0.01
0
15
0.785
0.785
0.77194
0.77598
0.785
0.000079389
334.19582
87607428
1456578
1.66262
Finished
andberg9
43m 19s
ssf
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.8,0.85]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='ssf' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
true
0.0005
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0005, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.01, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
target_modules=['linear', 'dense', 'batchnorm2d', 'conv2d'] adapter_type='ssf' task_type='SEQ_CLS' init_scale=1.0 init_shift=0.5
-
-
-
-
-
-
0.01
0
20
0.6225
0.6225
0.61456
0.62731
0.6225
0.00032725
329.56991
86394774
167524
0.19391
Finished
andberg9
17m 31s
none-classifier
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.8,0.85]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-classifier' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
true
0.0005
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0005, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.01, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=True gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
adapter_type='none-classifier' task_type='SEQ_CLS'
-
-
-
-
-
-
0.01
0
16
0.6125
0.6125
0.60605
0.64327
0.6125
0.00038396
328.93085
86227250
39986
0.046373
Finished
andberg9
43m 29s
adalora
["gaussian_noise","time_mask","polarity_inversion"]
1
8
[0.85,0.89]
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=10 pinned_memory=True epochs=20 save_model=False test_size=0.2 inference_size=0.0 val_size=0.1 sweep_count=200 accumulation_steps=1 patience=5 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='adalora' early_stopping=True checkpointing=True monitor='val_acc' mode='max' save_top_k=1 test_during_training=True test_during_training_freq=1 distributed_training=False num_gpus=2 strategy='ddp'
false
0.0005
-
-
-
-
50
optimizer_type='adamw' adam=AdamConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0005, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.1, amsgrad=False) warmup=WarmupConfig(enabled=False, warmup_steps=100, warmup_start_lr=1e-06, warmup_method='linear') scheduler_type='cosine_annealing_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='max', factor=0.85, patience=3, threshold=0.0001, threshold_mode='rel', cooldown=0, min_lr=0.0, eps=1e-08) step_lr=StepLRConfig(step_size=30, gamma=0.1) cosine_annealing_lr=CosineAnnealingLRConfig(T_max=50, eta_min=0.0) gradient_clipping_enabled=False gradient_clip_val=5.0 gradient_clip_algorithm='norm'
-
-
-
-
-
-
-
-
-
-
AdaLoraConfig(peft_type=<PeftType.ADALORA: 'ADALORA'>, auto_mapping=None, base_model_name_or_path='MIT/ast-finetuned-audioset-10-10-0.4593', revision=None, task_type='SEQ_CLS', inference_mode=False, r=8, target_modules=['audio_spectrogram_transformer.encoder.layer.0.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.0.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.0.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.0.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.0.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.1.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.1.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.1.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.1.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.2.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.2.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.2.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.2.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.3.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.3.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.3.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.3.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.4.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.4.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.4.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.4.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.5.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.5.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.5.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.5.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.6.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.6.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.6.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.6.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.7.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.7.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.7.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.7.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.8.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.8.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.8.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.8.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.9.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.9.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.9.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.9.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.10.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.10.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.10.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.10.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.query', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.key', 'audio_spectrogram_transformer.encoder.layer.11.attention.attention.value', 'audio_spectrogram_transformer.encoder.layer.11.attention.output.dense', 'audio_spectrogram_transformer.encoder.layer.11.intermediate.dense', 'audio_spectrogram_transformer.encoder.layer.11.output.dense', 'classifier.dense'], lora_alpha=4, lora_dropout=0.0, fan_in_fan_out=False, bias='none', use_rslora=False, modules_to_save=['classifier', 'score'], init_lora_weights=True, layers_to_transform=None, layers_pattern=None, rank_pattern=None, alpha_pattern={}, megatron_config=None, megatron_core='megatron.core', loftq_config={}, use_dora=False, layer_replication=None, runtime_config=LoraRuntimeConfig(ephemeral_gpu_offload=False), target_r=256, init_r=4, tinit=0, tfinal=0, deltaT=1, beta1=0.85, beta2=0.85, orth_reg_weight=0.5, total_step=None)
-
-
-
-
-
-
0.1
0
17
0.7975
0.7975
0.78313
0.79975
0.7975
0.00037044
331.64101
86937702
707103
0.81334
1-20
of 36