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

BandwiseBinaryClassifier is no longer work in progress

Browse Source
This commit is contained in:
Si11ium
2020-05-05 10:58:36 +02:00
parent 451f78f820
commit c2860b0aed
2 changed files with 29 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

46
models/bandwise_binary_classifier.py
View File

@ -13,11 +13,11 @@ from models.module_mixins import BaseOptimizerMixin, BaseTrainMixin, BaseValMixi
class BandwiseBinaryClassifier(BaseModuleMixin_Dataloaders,
BaseTrainMixin,
BaseValMixin,
BaseOptimizerMixin,
LightningBaseModule
):
BaseTrainMixin,
BaseValMixin,
BaseOptimizerMixin,
LightningBaseModule
):
def __init__(self, hparams):
super(BandwiseBinaryClassifier, self).__init__(hparams)
@ -25,7 +25,7 @@ class BandwiseBinaryClassifier(BaseModuleMixin_Dataloaders,
# Dataset and Dataloaders
# =============================================================================
# Transforms
transforms = Compose([AudioToMel(), MelToImage(), ToTensor(), NormalizeLocal()])
transforms = Compose([AudioToMel(n_mels=32), MelToImage(), ToTensor(), NormalizeLocal()])
# Datasets
from datasets.binar_masks import BinaryMasksDataset
self.dataset = Namespace(
@ -44,33 +44,35 @@ class BandwiseBinaryClassifier(BaseModuleMixin_Dataloaders,
self.criterion = nn.BCELoss()
self.n_band_sections = 5
# Utility Modules
# Modules
self.split = HorizontalSplitter(self.in_shape, self.n_band_sections)
self.conv_dict = ModuleDict()
# Modules with Parameters
modules = {f"conv_1_{band_section}":
ConvModule(self.in_shape, self.conv_filters[0], 3, conv_stride=2, **self.params.module_kwargs)
for band_section in range(self.n_band_sections)}
modules.update({f"conv_2_{band_section}":
ConvModule(self.conv_1.shape, self.conv_filters[1], 3, conv_stride=2,
self.conv_dict.update({f"conv_1_{band_section}":
ConvModule(self.split.shape, self.conv_filters[0], 3, conv_stride=1, **self.params.module_kwargs)
for band_section in range(self.n_band_sections)}
)
self.conv_dict.update({f"conv_2_{band_section}":
ConvModule(self.conv_dict['conv_1_1'].shape, self.conv_filters[1], 3, conv_stride=1,
**self.params.module_kwargs) for band_section in range(self.n_band_sections)}
)
modules.update({f"conv_3_{band_section}":
ConvModule(self.conv_2.shape, self.conv_filters[2], 3, conv_stride=2,
self.conv_dict.update({f"conv_3_{band_section}":
ConvModule(self.conv_dict['conv_2_1'].shape, self.conv_filters[2], 3, conv_stride=1,
**self.params.module_kwargs)
for band_section in range(self.n_band_sections)}
)
self.merge = HorizontalMerger(self.conv_dict['conv_3_1'].shape, self.n_band_sections)
self.flat = Flatten(self.merge.shape)
self.full_1 = nn.Linear(self.flat.shape, self.params.lat_dim, self.params.bias)
self.full_2 = nn.Linear(self.full_1.out_features, self.full_1.out_features // 2, self.params.bias)
self.full_out = nn.Linear(self.full_2.out_features, 1, self.params.bias)
# Utility Modules
self.merge = HorizontalMerger(self.split.shape, self.n_band_sections)
self.conv_dict = ModuleDict(modules=modules)
self.flat = Flatten(self.conv_3.shape)
self.dropout = nn.Dropout2d(self.params.dropout) if self.params.dropout else lambda x: x
self.activation = self.params.activation()
self.sigmoid = nn.Sigmoid()
@ -78,11 +80,11 @@ class BandwiseBinaryClassifier(BaseModuleMixin_Dataloaders,
def forward(self, batch, **kwargs):
tensors = self.split(batch)
for idx, tensor in enumerate(tensors):
tensor[idx] = self.conv_dict[f"conv_1_{idx}"](tensor)
tensors[idx] = self.conv_dict[f"conv_1_{idx}"](tensor)
for idx, tensor in enumerate(tensors):
tensor[idx] = self.conv_dict[f"conv_2_{idx}"](tensor)
tensors[idx] = self.conv_dict[f"conv_2_{idx}"](tensor)
for idx, tensor in enumerate(tensors):
tensor[idx] = self.conv_dict[f"conv_3_{idx}"](tensor)
tensors[idx] = self.conv_dict[f"conv_3_{idx}"](tensor)
tensor = self.merge(tensors)
tensor = self.flat(tensor)