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

bugs fixed, binary datasets working

Browse Source
This commit is contained in:
Steffen
2021-03-27 18:23:51 +01:00
parent 37e36df0a8
commit 76eb567eed
4 changed files with 33 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files

45
util/module_mixins.py
View File

@@ -96,39 +96,27 @@ class ValMixin:
for file_name in sorted_y:
sorted_y.update({file_name: torch.stack(sorted_y[file_name])})
#y_mean = torch.stack(
# [torch.mean(x, dim=0, keepdim=True) if x.shape[0] > 1 else x for x in sorted_y.values()]
#).squeeze()
#if y_mean.ndim == 1:
# y_mean = y_mean.unsqueeze(0)
#if sorted_batch_y.ndim == 1:
# sorted_batch_y = sorted_batch_y.unsqueeze(-1)
#
#mean_vote_loss = self.ce_loss(y_mean, sorted_batch_y)
#summary_dict.update(val_mean_vote_loss=mean_vote_loss)
if self.params.n_classes <= 2:
mean_sorted_y = torch.stack([x.mean(dim=0) if x.shape[0] > 1 else x for x in sorted_y.values()])
mean_sorted_y = torch.stack([x.mean(dim=0) if x.shape[0] > 1 else x for x in sorted_y.values()]).squeeze()
max_vote_loss = self.bce_loss(mean_sorted_y.float(), sorted_batch_y.float())
# Sklearn Scores
additional_scores = self.additional_scores(dict(y=mean_sorted_y, batch_y=sorted_batch_y))
else:
y_max = torch.stack(
[torch.argmax(x.mean(dim=0)) if x.shape[0] > 1 else torch.argmax(x) for x in sorted_y.values()]
).squeeze()
y_one_hot = torch.nn.functional.one_hot(y_max, num_classes=self.params.n_classes).float()
max_vote_loss = self.ce_loss(y_one_hot, sorted_batch_y)
summary_dict.update(val_max_vote_loss=max_vote_loss)
# Sklearn Scores
additional_scores = self.additional_scores(dict(y=y_one_hot, batch_y=sorted_batch_y))
summary_dict.update(val_max_vote_loss=max_vote_loss, **additional_scores)
summary_dict.update({f'mean_{key}': torch.mean(torch.stack([output[key]
for output in outputs]))
for key in keys if 'loss' in key}
)
# Sklearn Scores
if self.params.n_classes <= 2:
additional_scores = self.additional_scores(dict(y=y_max, batch_y=sorted_batch_y))
else:
additional_scores = self.additional_scores(dict(y=y_one_hot, batch_y=sorted_batch_y))
summary_dict.update(**additional_scores)
pl_metrics, pl_images = self.metrics.compute_and_prepare()
self.metrics.reset()
@@ -166,19 +154,20 @@ class TestMixin:
for file_name in sorted_y:
sorted_y.update({file_name: torch.stack(sorted_y[file_name])})
y_max = torch.stack(
[torch.argmax(x.mean(dim=0)) if x.shape[0] > 1 else torch.argmax(x) for x in sorted_y.values()]
).squeeze().cpu()
if self.params.n_classes == 5:
if self.params.n_classes > 2:
pred = torch.stack(
[torch.argmax(x.mean(dim=0)) if x.shape[0] > 1 else torch.argmax(x) for x in sorted_y.values()]
).squeeze().cpu()
class_names = {val: key for val, key in
enumerate(['background', 'chimpanze', 'geunon', 'mandrille', 'redcap'])}
elif self.params.n_classes == 2:
class_names = {val: key for val, key in ['negative', 'positive']}
else:
raise AttributeError(f'n_classes has to be any of: [2, 5]')
pred = torch.stack([x.mean(dim=0) if x.shape[0] > 1 else x for x in sorted_y.values()]).squeeze()
class_names = {val: key for val, key in ['negative', 'positive']}
df = pd.DataFrame(data=dict(filename=[Path(x).name for x in sorted_y.keys()],
prediction=[class_names[x.item()] for x in y_max.cpu()]))
prediction=[class_names[x.item()] for x in pred.cpu()]))
result_file = Path(self.logger.log_dir / 'predictions.csv')
if result_file.exists():
try: