2c700450b5
Change Embedding's input_dim argument
...
Wrong argument for the decoder's embedding layer.
2020-10-20 12:31:25 -04:00
e0cae0c7be
Replace deprecated method
...
See https://www.tensorflow.org/api_docs/python/tf/keras/Sequential?hl=en#predict_classes .
2020-10-20 12:31:25 -04:00
a2ffc37d2f
Modify creation of possible char list
...
I concatenated the string of all digits (+ comma and space) to the argument of function sorted ∘ set. Also, the digit '0' was written twice in the digit string.
2020-10-20 12:31:04 -04:00
7848437dc2
Correct typo
2020-10-19 12:44:33 -04:00
9d99ae9f9f
Correct small coding typo
2020-10-19 12:38:17 -04:00
8b00e3130d
Solution 9 version 4 requires TF 2.2 (add warning)
2020-04-22 19:59:35 +12:00
1f0bbc782a
Add more examples of TF-Addons seq2seq usage
2020-04-22 19:21:56 +12:00
ffd127eb7a
Add solutions to the coding exercises: Embedded Reber grammars, Encoder-Decoder and text generation using the transformers library
2020-04-19 16:01:14 +12:00
2f7ab70295
Update notebooks to latest nbformat
2020-04-06 19:13:12 +12:00
ba6c75d6e5
Update notebook 16 to TF 2.1
2020-04-06 15:22:22 +12:00
88dccccd5f
Make notebooks 14 to 19 runnable in Colab without changes
2019-11-06 21:06:55 +08:00
22aa520ccc
issue 32: remove dropout in stateful RNN
2019-09-28 12:06:05 -04:00
f0f83903bd
Fix the transformer (use final encoder outputs)
2019-05-10 21:30:18 +08:00
6f98bc5a04
Stateful RNNs now support recurrent_dropout
2019-04-21 15:55:39 +08:00
df9b91e2e8
Rename chapters 11 to 15 and split chapter 15 into 15 and 16
2019-04-16 20:39:14 +08:00
Ian Beauregard