Andberg9's workspace
Runs
144
Name
144 visualized
val_acc
best_val_accuracy
0.33387
-
0.53065
0
0.97419
0
0.64355
0
0.59677
0
0.94516
0
0.45161
0
0.49355
0
0.94516
0
0.3425
0
0.585
0
0.5025
0
0.5075
0
0.5325
0
0.4775
0
0.35
0
0.52
0
0.475
0
0.64376
0
0.58076
0
State
Notes
User
Tags
Created
Runtime
Sweep
augmentations.augmentations
augmentations.augmentations_per_sample
general_config
loss.label_smoothing
loss_config
method
metric.goal
metric.name
name
num_classes
optimizer.adamw.lr
optimizer_config
parameters.loss.label_smoothings
parameters.optimizer.adamw.lrs
peft_config
peft_scheduling_config
project
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
total_train_time
train_acc
train_f1
train_loss
train_precision
train_recall
trainer/global_step
val_f1
val_loss
val_precision
val_recall
Crashed
-
andberg9
30m 31s
["time_mask","time_stretch","sin_distortion"]
0
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.2
type='cross_entropy' label_smoothing=0.2 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
31
1e-2
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.01, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
15
-
-
-
-
-
0.01
328.87512
86212639
86212639
100
-
0.32823
0.30578
2.71287
0.29931
0.32823
4960
0.27419
2.71402
0.34619
0.33387
Finished
-
andberg9
39m 28s
["time_mask","time_stretch","sin_distortion"]
0
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.2
type='cross_entropy' label_smoothing=0.2 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
31
1e-3
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
21
-
-
-
-
-
0.001
328.87512
86212639
86212639
100
39:23
0.56815
0.55531
2.19067
0.55237
0.56815
6819
0.49565
2.2027
0.52781
0.53065
Finished
-
andberg9
1h 11m 15s
["time_mask","time_stretch","sin_distortion"]
0
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.2
type='cross_entropy' label_smoothing=0.2 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
31
1e-4
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
39
-
-
-
-
-
0.00001
328.87512
86212639
86212639
100
71:09
1
1
1.14976
1
1
12399
0.97453
1.21654
0.97585
0.97419
Finished
-
andberg9
2h 39m 45s
["time_mask","time_stretch","sin_distortion"]
0
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.1
type='cross_entropy' label_smoothing=0.1 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
31
1e-2
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.01, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
89
-
-
-
-
-
0.00001
328.87512
86212639
86212639
100
159:39
0.69919
0.68107
1.58435
0.68382
0.69919
27899
0.62879
1.69491
0.63691
0.64355
Finished
-
andberg9
48m 51s
["time_mask","time_stretch","sin_distortion"]
0
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.1
type='cross_entropy' label_smoothing=0.1 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
31
1e-3
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
26
-
-
-
-
-
0.001
328.87512
86212639
86212639
100
48:45
0.72823
0.72386
1.4651
0.72421
0.72823
8369
0.5762
1.77294
0.65867
0.59677
Finished
-
andberg9
37m 37s
["time_mask","time_stretch","sin_distortion"]
0
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.1
type='cross_entropy' label_smoothing=0.1 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
31
1e-4
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
20
-
-
-
-
-
0.0001
328.87512
86212639
86212639
100
37:31
0.99556
0.99558
0.6692
0.9957
0.99556
6509
0.94559
0.81449
0.9497
0.94516
Finished
-
andberg9
2h 39m 26s
["time_mask","time_stretch","sin_distortion"]
0
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0
type='cross_entropy' label_smoothing=0.0 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
31
1e-2
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.01, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
89
-
-
-
-
-
0.00001
328.87512
86212639
86212639
100
159:21
0.48669
0.46203
1.7178
0.45634
0.48669
27899
0.43217
1.807
0.42856
0.45161
Finished
-
andberg9
41m 37s
["time_mask","time_stretch","sin_distortion"]
0
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0
type='cross_entropy' label_smoothing=0.0 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
31
1e-3
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
22
-
-
-
-
-
0.001
328.87512
86212639
86212639
100
41:32
0.46653
0.44429
1.67549
0.45314
0.46653
7129
0.46631
1.53738
0.53081
0.49355
Finished
-
andberg9
41m 14s
["time_mask","time_stretch","sin_distortion"]
0
model_type='ast' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=False test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0
type='cross_entropy' label_smoothing=0.0 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
31
1e-4
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
22
-
-
-
-
-
0.0001
328.87512
86212639
86212639
100
41:08
0.99234
0.99234
0.026871
0.99236
0.99234
7129
0.94415
0.21972
0.94886
0.94516
Finished
-
andberg9
12m 16s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.2
type='cross_entropy' label_smoothing=0.2 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
50
1e-2
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.01, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
71
0.3425
0.3425
0.30936
0.30814
0.3425
0.0001
67.28135
17637402
17637402
100
12:10
0.3825
0.36186
2.83994
0.35992
0.3825
14399
0.30936
2.94722
0.30814
0.3425
Finished
-
andberg9
9m 39s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.2
type='cross_entropy' label_smoothing=0.2 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
50
1e-3
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
55
0.585
0.585
0.59402
0.63865
0.585
0.0001
67.28135
17637402
17637402
100
09:32
0.96187
0.96185
1.58106
0.96283
0.96187
11199
0.59402
2.42682
0.63865
0.585
Finished
-
andberg9
13m 32s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.2
type='cross_entropy' label_smoothing=0.2 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
50
1e-4
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
79
0.5025
0.5025
0.48862
0.51971
0.5025
0.000001
67.28135
17637402
17637402
100
13:26
0.97062
0.97064
1.45792
0.97137
0.97062
15999
0.48862
2.74831
0.51971
0.5025
Finished
-
andberg9
9m 32s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.1
type='cross_entropy' label_smoothing=0.1 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
50
1e-2
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.01, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
55
0.5075
0.5075
0.49748
0.56392
0.5075
0.001
67.28135
17637402
17637402
100
09:25
0.7075
0.70286
1.69458
0.70666
0.7075
11199
0.49748
2.40916
0.56392
0.5075
Finished
-
andberg9
8m 13s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.1
type='cross_entropy' label_smoothing=0.1 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
50
1e-3
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
47
0.5325
0.5325
0.53665
0.60704
0.5325
0.0001
67.28135
17637402
17637402
100
08:06
0.93563
0.93531
1.10309
0.93736
0.93563
9599
0.53665
8.80455
0.60704
0.5325
Finished
-
andberg9
7m 56s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.1
type='cross_entropy' label_smoothing=0.1 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
50
1e-4
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
45
0.4775
0.4775
0.46734
0.48632
0.4775
0.00001
67.28135
17637402
17637402
100
07:49
0.9475
0.94753
0.99227
0.94926
0.9475
9199
0.46734
2.49942
0.48632
0.4775
Finished
-
andberg9
15m 14s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0
type='cross_entropy' label_smoothing=0.0 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
50
1e-2
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.01, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
89
0.35
0.35
0.34198
0.37278
0.35
0.00001
67.28135
17637402
17637402
100
15:06
0.37438
0.36012
2.06874
0.35707
0.37438
17999
0.34198
43.31335
0.37278
0.35
Finished
-
andberg9
12m 2s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0
type='cross_entropy' label_smoothing=0.0 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
50
1e-3
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
70
0.52
0.52
0.52226
0.56423
0.52
0.00001
67.28135
17637402
17637402
100
11:54
0.92688
0.92674
0.25764
0.92833
0.92688
14199
0.52226
1.97529
0.56423
0.52
Finished
-
andberg9
12m 32s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0
type='cross_entropy' label_smoothing=0.0 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
50
1e-4
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.0001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
74
0.475
0.475
0.46175
0.47896
0.475
0.000001
67.28135
17637402
17637402
100
12:26
0.95063
0.95058
0.22652
0.95183
0.95063
14999
0.46175
2.69097
0.47896
0.475
Finished
-
andberg9
39m 44s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.2
type='cross_entropy' label_smoothing=0.2 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
10
1e-2
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.01, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
54
0.64376
0.64376
0.62774
0.6538
0.64376
0.001
67.00776
17565682
17565682
100
38:53
0.99478
0.99478
0.89969
0.99479
0.99478
54064
0.62774
1.74561
0.6538
0.64376
Finished
-
andberg9
19m 18s
["time_mask","time_stretch","sin_distortion"]
0
model_type='efficientnet_b4' save_dataloader=False batch_size=8 seed=42 num_cuda_workers=5 pinned_memory=True epochs=90 save_model=False from_scratch=True test_size=0.0 inference_size=0.0 val_size=0.2 sweep_count=9 accumulation_steps=1 patience=15 use_wandb=True use_sweep=True torch_viz=False use_kfold=False k_folds=5 adapter_type='none-full' compute_fisher=False fisher_mc_samples=100 save_fim_heatmap=True save_fim_epochs=True 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'
0.2
type='cross_entropy' label_smoothing=0.2 class_weights=None focal_alpha=None focal_gamma=2.0
-
-
-
-
10
1e-3
optimizer_type='adamw' adam=AdamConfig(lr=1e-08, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.0, amsgrad=False) adamw=AdamWConfig(lr=0.001, betas=(0.99, 0.999), eps=1e-08, weight_decay=0.001, amsgrad=False) adamspd=AdamSPDConfig(lr=1e-08, 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='step_lr' reduce_lr_on_plateau=ReduceLROnPlateauConfig(mode='min', factor=0.1, patience=10, 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='none-full' task_type='SEQ_CLS'
enabled=False model_name=None schedule=[PEFTScheduleStep(start_epoch=1, peft_method='none-classifier', merge_previous=True), PEFTScheduleStep(start_epoch=2, peft_method='none-full', merge_previous=True), PEFTScheduleStep(start_epoch=3, peft_method='ia3', merge_previous=True)] auto_merge=True
-
25
0.58076
0.58076
0.5611
0.57303
0.58076
0.001
67.00776
17565682
17565682
100
18:27
0.96145
0.96143
0.96617
0.96145
0.96145
25557
0.5611
1.83454
0.57303
0.58076
1-20
of 144