user avatar
user avatar
user avatar
hH1sG0n3
  • Member for 1 year, 11 months
  • Last seen this week
18 votes
4 answers
4k views

In a Transformer model, why does one sum positional encoding to the embedding rather than concatenate it?

4 votes
1 answer
764 views

Self-attention mechanism did not improve the LSTM classification model

  • 115
4 votes
1 answer
975 views

Train LSTM model with multiple time series

  • 41
4 votes
2 answers
742 views

In sequence models, is it possible to have training batches with different timesteps each to reduce the required padding per input sequence?

  • 63
3 votes
2 answers
765 views

What is the relation between input into LSTM and number of cells?

  • 85
3 votes
2 answers
895 views

Why is the "dying ReLU" problem not present in most modern deep learning architectures?

  • 121
2 votes
1 answer
235 views

Does multi-head attention remove the need for self-attention?

2 votes
1 answer
370 views

A simple attention based text prediction model from scratch using pytorch

  • 271
1 vote
1 answer
343 views

Classifying videos with varying length using ConvLSTM2D in tensorflow

1 vote
1 answer
15 views

Structuring extensive medical histories + demographic information for prediciting future medical outcomes

  • 111
1 vote
0 answers
417 views

What is the reason behind Keras choice of default (recurrent) activation functions in LSTM networks