User user3668129

6 votes

1 answer

785 views

Why autoencoders use binary_crossentropy loss and not mean squared error?

asked Jun 29, 2021 at 10:25

4 votes

1 answer

213 views

How FLAIR NER algorithm detects entities with typo?

asked Feb 5, 2022 at 8:05

3 votes

1 answer

5k views

How batch normalization layer resolve the vanishing gradient problem?

asked Jun 2, 2021 at 4:40

2 votes

0 answers

114 views

Convolutional autoencoder - why keras example is asymmetry model?

asked Jun 22, 2021 at 10:39

1 vote

0 answers

18 views

Why FLAIR does't recognize the entire location name

asked Feb 5, 2022 at 7:55

1 vote

1 answer

114 views

In LSTM why h_t output twice?

asked Jul 2, 2022 at 16:25

1 vote

2 answers

48 views

What we can learn from the data if PCA scree plot bins are almost the same?

asked Jun 13, 2021 at 12:52

1 vote

2 answers

1k views

T-SNE with high number of features

asked Jun 22, 2021 at 6:01

1 vote

0 answers

194 views

If the input to the autoencoder is normalized, do we need to use sigmoid on the last layer?

asked Jul 1, 2021 at 18:27

1 vote

0 answers

27 views

Is there a reason not to wirk with AMP (automatic mixed precision)?

asked Dec 5, 2021 at 12:29

1 vote

1 answer

54 views

Why VAE Encoder outputs log variance and not standard deviation?

asked Dec 26, 2022 at 13:50

1 vote

1 answer

513 views

Does settings $\beta_1 = 0$ or $\beta_2 = 0$ means that ADAM behaves as RMSprop or Momentum?

asked Dec 13, 2022 at 12:52

1 vote

0 answers

200 views

How Adam optimizer influence the learning rate? [closed]

asked Dec 17, 2022 at 15:08

0 votes

0 answers

10 views

In RMSProp how squares of the gradients dumps out the oscillations?

asked Dec 20, 2022 at 12:20

0 votes

1 answer

79 views

What does the ellipse in Gradient descent describes?

asked Dec 21, 2022 at 9:22

0 votes

0 answers

19 views

What is the meaning of uncentered variance?

asked Dec 22, 2022 at 8:52

0 votes

0 answers

20 views

Do linear layer after GRU saved the sequence output order?

asked Dec 29, 2022 at 13:47

0 votes

1 answer

36 views

The meaning of P and degree of freedom in T-Test

asked Jan 26 at 11:07

0 votes

0 answers

15 views

When using wav2vec2 inference, does the quantitation being used?

asked Feb 21 at 8:29

0 votes

0 answers

52 views

Which decoder method is used with “facebook/wav2vec2-base-960h”?

asked Mar 1 at 16:32

0 votes

1 answer

83 views

Are there any pre-trained non english model of deepspeech?

asked Mar 4 at 7:19

0 votes

0 answers

170 views

Which loss function is used on Whisper model?

asked Mar 9 at 5:48

0 votes

0 answers

21 views

How the embedding model (x-vectors) trained?

asked Mar 30 at 7:44

0 votes

0 answers

33 views

Why speaker diarization DER on Ego4D dataset is so bad?

asked Mar 30 at 7:58

0 votes

1 answer

39 views

What is the difference between VAD and Speaker Segmentation?

asked Apr 3 at 14:50

0 votes

0 answers

18 views

Can we use the pretrained WavLM on Portuguese?

asked Sep 20 at 8:47

0 votes

0 answers

6 views

How to get / calculate statistics about voxceleb2?

asked Nov 30 at 8:30

0 votes

1 answer

1k views

What are the differences between bert embedding and flair embedding

asked Feb 1, 2022 at 11:24

0 votes

1 answer

67 views

reprocessing steps for images before training classification models

asked May 10, 2021 at 8:42

0 votes

1 answer

22 views

Using vgg16 or inception with wights equals to None

asked May 11, 2021 at 8:34

Stack Exchange Network

35 Questions

Why autoencoders use binary_crossentropy loss and not mean squared error?

How FLAIR NER algorithm detects entities with typo?

How batch normalization layer resolve the vanishing gradient problem?

Convolutional autoencoder - why keras example is asymmetry model?

Why FLAIR does't recognize the entire location name

In LSTM why h_t output twice?

What we can learn from the data if PCA scree plot bins are almost the same?

T-SNE with high number of features

If the input to the autoencoder is normalized, do we need to use sigmoid on the last layer?

Is there a reason not to wirk with AMP (automatic mixed precision)?

Why VAE Encoder outputs log variance and not standard deviation?

Does settings $\beta_1 = 0$ or $\beta_2 = 0$ means that ADAM behaves as RMSprop or Momentum?

How Adam optimizer influence the learning rate? [closed]

In RMSProp how squares of the gradients dumps out the oscillations?

What does the ellipse in Gradient descent describes?

What is the meaning of uncentered variance?

Do linear layer after GRU saved the sequence output order?

The meaning of P and degree of freedom in T-Test

When using wav2vec2 inference, does the quantitation being used?

Which decoder method is used with “facebook/wav2vec2-base-960h”?

Are there any pre-trained non english model of deepspeech?

Which loss function is used on Whisper model?

How the embedding model (x-vectors) trained?

Why speaker diarization DER on Ego4D dataset is so bad?

What is the difference between VAD and Speaker Segmentation?

Can we use the pretrained WavLM on Portuguese?

How to get / calculate statistics about voxceleb2?

What are the differences between bert embedding and flair embedding

reprocessing steps for images before training classification models

Using vgg16 or inception with wights equals to None