Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jun 1, 2020

• remove unused import

  bentrevett

  committed
  d876a1d

  Copy full SHA for d876a1d

Commits on Apr 17, 2020

• Merge pull request #100 from rakeshchada/master

  Show description for 6559ece
  bentrevett

  authored
  6559ece

  Copy full SHA for 6559ece

• Merge pull request #99 from HelgeS/patch-1

  Show description for 64c6f75
  bentrevett

  authored
  64c6f75

  Copy full SHA for 64c6f75

Commits on Apr 16, 2020

• fix: use separate LayerNorm layers for attention and feedforward blocks

  rakeshchada

  committed
  0f7a00a

  Copy full SHA for 0f7a00a

• Fixed Colab links for notebooks 5 & 6

  HelgeS

  authored
  f6a4d76

  Copy full SHA for f6a4d76

• (#97) fixed typo on encoder gru equation

  bentrevett

  committed
  49cbdd3

  Copy full SHA for 49cbdd3

Commits on Apr 12, 2020

• fixed unsqueeze (#96)

  bentrevett

  committed
  041994c

  Copy full SHA for 041994c

Commits on Mar 23, 2020

• (#86) fixed typo

  bentrevett

  committed
  e8209a7

  Copy full SHA for e8209a7

• (#87) fixed typos

  bentrevett

  committed
  523f810

  Copy full SHA for 523f810

Commits on Mar 12, 2020

• (#83) fixed comment typos and added attention x v matmul tensor sizes.

  bentrevett

  committed
  d718eb7

  Copy full SHA for d718eb7

Commits on Feb 25, 2020

• (#79) fixed src_len tensor size in encoder

  bentrevett

  committed
  48996d5

  Copy full SHA for 48996d5

Commits on Feb 3, 2020

• updated notebooks to reflect change of v from tensor to linear layer

  bentrevett

  committed
  dce9b70

  Copy full SHA for dce9b70

• Merge pull request #75 from M-e-r-c-u-r-y/master

  Show description for 8b0b080
  bentrevett

  authored
  8b0b080

  Copy full SHA for 8b0b080

Commits on Feb 2, 2020

• Updated Attention Class to make v a Linear layer to simplify code

  M-e-r-c-u-r-y

  committed
  9d03d19

  Copy full SHA for 9d03d19

• Updated Attention Class to make v a Linear layer to simplify code

  M-e-r-c-u-r-y

  committed
  4f47fd2

  Copy full SHA for 4f47fd2

Commits on Jan 28, 2020

• no longer passing modules/layers as arguments to transformer

  bentrevett

  committed
  e72ae20

  Copy full SHA for e72ae20

Commits on Jan 22, 2020

• changed v initialization from rand to randn. added some more description about v

  bentrevett

  committed
  5f19500

  Copy full SHA for 5f19500

• (#71) typo on convseq2seq intro

  bentrevett

  committed
  dfa0176

  Copy full SHA for dfa0176

Commits on Jan 21, 2020

• max length for positional encoding is now an optional parameter

  bentrevett

  committed
  7f583f3

  Copy full SHA for 7f583f3

• max length for positional encoding is now an optional parameter

  bentrevett

  committed
  d7316f6

  Copy full SHA for d7316f6

• clarified difference between tokens and embeddings of tokens

  bentrevett

  committed
  1df5de4

  Copy full SHA for 1df5de4

• clarified difference between tokens and embeddings of tokens

  bentrevett

  committed
  b9d5c90

  Copy full SHA for b9d5c90

• clarified difference between tokens and embeddings of tokens

  bentrevett

  committed
  b88f6d4

  Copy full SHA for b88f6d4

• updated required pytorch and torchtext

  bentrevett

  committed
  dee4844

  Copy full SHA for dee4844

• updated explanations for color change in the backward gru images

  bentrevett

  committed
  2bc44c5

  Copy full SHA for 2bc44c5

• re-ran model and updated inference explanations

  bentrevett

  committed
  b2a0e71

  Copy full SHA for b2a0e71

• updated bleu score comment

  bentrevett

  committed
  90cc746

  Copy full SHA for 90cc746

• changed 'you' to remove personal text

  bentrevett

  committed
  c590ffc

  Copy full SHA for c590ffc

• mentioned words are actually embedded words. change i/you to we/us

  bentrevett

  committed
  6214a2a

  Copy full SHA for 6214a2a

• mentioned how words are actually embedded

  bentrevett

  committed
  be6df75

  Copy full SHA for be6df75

• updated assets/images to include embedding layer

  bentrevett

  committed
  a1bf0d4

  Copy full SHA for a1bf0d4

• removed bleu file as now using torchtext.metrics bleu

  bentrevett

  committed
  d136169

  Copy full SHA for d136169

• added official torchtext bleu implementation

  bentrevett

  committed
  a18c7ca

  Copy full SHA for a18c7ca

• reran with pytorch 1.4 and torchtext 0.5

  bentrevett

  committed
  9c81bc4

  Copy full SHA for 9c81bc4

Commits on Jan 13, 2020

• fixed colab links

  bentrevett

  committed
  7df6989

  Copy full SHA for 7df6989

Pagination

Previous

Next