steffen/Mel_Vision_Transformer_ComParE-21
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

CCS intergration dataloader

Browse Source
This commit is contained in:
Steffen 2021-03-19 18:05:17 +01:00
parent d30edbda6e
commit 78b3139d1a
2 changed files with 21 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
multi_run.py
View File

@ -10,27 +10,30 @@ import itertools
if __name__ == '__main__': if __name__ == '__main__':
# Set new values # Set new values
hparams_dict = dict(seed=[69], hparams_dict = dict(seed=range(10),
model_name=['VisualTransformer'], model_name=['VisualTransformer'],
data_name=['CCSLibrosaDatamodule'], data_name=['CCSLibrosaDatamodule'],
batch_size=[5], batch_size=[50],
max_epochs=[200], max_epochs=[200],
variable_length=[False],
sample_segment_len=[40],
sample_hop_len=[15],
random_apply_chance=[0.5], # trial.suggest_float('random_apply_chance', 0.1, 0.5, step=0.1), random_apply_chance=[0.5], # trial.suggest_float('random_apply_chance', 0.1, 0.5, step=0.1),
loudness_ratio=[0.3], # trial.suggest_float('loudness_ratio', 0.0, 0.5, step=0.1), loudness_ratio=[0], # trial.suggest_float('loudness_ratio', 0.0, 0.5, step=0.1),
shift_ratio=[0.3], # trial.suggest_float('shift_ratio', 0.0, 0.5, step=0.1), shift_ratio=[0.3], # trial.suggest_float('shift_ratio', 0.0, 0.5, step=0.1),
noise_ratio=[0.3], # trial.suggest_float('noise_ratio', 0.0, 0.5, step=0.1), noise_ratio=[0.3], # trial.suggest_float('noise_ratio', 0.0, 0.5, step=0.1),
mask_ratio=[0.3], # trial.suggest_float('mask_ratio', 0.0, 0.5, step=0.1), mask_ratio=[0.3], # trial.suggest_float('mask_ratio', 0.0, 0.5, step=0.1),
lr=[1e-2], # trial.suggest_uniform('lr', 1e-3, 3e-3), lr=[1e-3], # trial.suggest_uniform('lr', 1e-3, 3e-3),
dropout=[0.2], # trial.suggest_float('dropout', 0.0, 0.3, step=0.05), dropout=[0.2], # trial.suggest_float('dropout', 0.0, 0.3, step=0.05),
lat_dim=[48], # 2 ** trial.suggest_int('lat_dim', 1, 5, step=1), lat_dim=[32], # 2 ** trial.suggest_int('lat_dim', 1, 5, step=1),
mlp_dim=[30], # 2 ** trial.suggest_int('mlp_dim', 1, 5, step=1), mlp_dim=[16], # 2 ** trial.suggest_int('mlp_dim', 1, 5, step=1),
head_dim=[12], # 2 ** trial.suggest_int('head_dim', 1, 5, step=1), head_dim=[6], # 2 ** trial.suggest_int('head_dim', 1, 5, step=1),
patch_size=[12], # trial.suggest_int('patch_size', 6, 12, step=3), patch_size=[12], # trial.suggest_int('patch_size', 6, 12, step=3),
attn_depth=[12], # trial.suggest_int('attn_depth', 2, 14, step=4), attn_depth=[12], # trial.suggest_int('attn_depth', 2, 14, step=4),
heads=[12], # trial.suggest_int('heads', 2, 16, step=2), heads=[6], # trial.suggest_int('heads', 2, 16, step=2),
scheduler=['LambdaLR'], # trial.suggest_categorical('scheduler', [None, 'LambdaLR']), scheduler=['LambdaLR'], # trial.suggest_categorical('scheduler', [None, 'LambdaLR']),
lr_scheduler_parameter=[0.95], # [0.98], lr_scheduler_parameter=[0.95], # [0.98],
embedding_size=[64], # trial.suggest_int('embedding_size', 12, 64, step=12), embedding_size=[30], # trial.suggest_int('embedding_size', 12, 64, step=12),
loss=['ce_loss'], loss=['ce_loss'],
sampler=['WeightedRandomSampler'], sampler=['WeightedRandomSampler'],
# rial.suggest_categorical('sampler', [None, 'WeightedRandomSampler']), # rial.suggest_categorical('sampler', [None, 'WeightedRandomSampler']),
@ -41,7 +44,7 @@ if __name__ == '__main__':
permutations_dicts = [dict(zip(keys, v)) for v in itertools.product(*values)] permutations_dicts = [dict(zip(keys, v)) for v in itertools.product(*values)]
for permutations_dict in tqdm(permutations_dicts, total=len(permutations_dicts)): for permutations_dict in tqdm(permutations_dicts, total=len(permutations_dicts)):
# Parse comandline args, read config and get model # Parse comandline args, read config and get model
cmd_args, *data_model_seed = parse_comandline_args_add_defaults( cmd_args, found_data_class, found_model_class, found_seed = parse_comandline_args_add_defaults(
'_parameters.ini', overrides=permutations_dict) '_parameters.ini', overrides=permutations_dict)
hparams = dict(**cmd_args) hparams = dict(**cmd_args)
@ -51,6 +54,6 @@ if __name__ == '__main__':
# RUN # RUN
# --------------------------------------- # ---------------------------------------
print(f'Running Loop, parameters are: {permutations_dict}') print(f'Running Loop, parameters are: {permutations_dict}')
run_lightning_loop(hparams, *data_model_seed) run_lightning_loop(hparams, found_data_class, found_model_class, seed=found_seed)
print(f'Done, parameters were: {permutations_dict}') print(f'Done, parameters were: {permutations_dict}')
pass pass

8
util/module_mixins.py
View File

@ -59,7 +59,13 @@ class ValMixin:
[torch.argmax(x.mean(dim=0)) if x.shape[0] > 1 else torch.argmax(x) for x in sorted_y.values()] [torch.argmax(x.mean(dim=0)) if x.shape[0] > 1 else torch.argmax(x) for x in sorted_y.values()]
).squeeze() ).squeeze()
y_one_hot = torch.nn.functional.one_hot(y_max, num_classes=self.params.n_classes).float() y_one_hot = torch.nn.functional.one_hot(y_max, num_classes=self.params.n_classes).float()
self.metrics.update(y_one_hot, torch.stack(tuple(sorted_batch_y.values())).long()) target_y = torch.stack(tuple(sorted_batch_y.values())).long()
if y_one_hot.ndim == 1:
y_one_hot = y_one_hot.unsqueeze(0)
if target_y.ndim == 1:
target_y = target_y.unsqueeze(0)
self.metrics.update(y_one_hot, target_y)
val_loss = self.ce_loss(y, batch_y.long()) val_loss = self.ce_loss(y, batch_y.long())