Aurélien Geron
|
2913a17508
|
Work around "import tensorflow.train" bug.
|
2019-10-12 17:42:41 +09:30 |
|
Aurélien Geron
|
5c9957065b
|
Work around "import tensorflow.train" bug.
|
2019-10-12 17:38:40 +09:30 |
|
Aurélien Geron
|
02205e034f
|
Remove from __future__ imports as we move away from Python 2
|
2019-10-12 17:07:53 +09:30 |
|
Aurélien Geron
|
7d0052f146
|
Import urllib directly instead of from six.moves, as we move away from Python 2
|
2019-10-12 16:42:23 +09:30 |
|
Aurélien Geron
|
f6dfa0ff76
|
Remove from __future__ imports as we move away from Python 2
|
2019-10-12 16:40:05 +09:30 |
|
Aurélien Geron
|
936e2cf50f
|
Import urllib directly instead of from six.moves, as we move away from Python 2
|
2019-10-12 16:30:13 +09:30 |
|
Aurélien Geron
|
07b83a5e09
|
Import urllib directly instead of from six.moves, as we move away from Python 2
|
2019-10-12 16:29:54 +09:30 |
|
Aurélien Geron
|
f5f1ce1ca0
|
Remove deepnote note about startup speed
|
2019-10-01 09:51:50 +09:30 |
|
Aurélien Geron
|
3f634d9eb9
|
Merge pull request #35 from Jirka-Lhotka/patch-1
Remove outdated note in Deepnote's description
|
2019-10-01 09:48:49 +09:30 |
|
Jirka-Lhotka
|
41686e16d2
|
Update Deepnote's note
|
2019-09-30 13:58:49 +02:00 |
|
Aurélien Geron
|
f798a5c932
|
Merge pull request #34 from thierryherrmann/no_dropout_in_stateful_rnn
issue 32: temporarily remove dropout in stateful RNN
|
2019-09-30 15:30:10 +09:30 |
|
Thierry Herrmann
|
22aa520ccc
|
issue 32: remove dropout in stateful RNN
|
2019-09-28 12:06:05 -04:00 |
|
Aurélien Geron
|
b9509ced94
|
First to do => First thing to do
|
2019-09-28 22:55:02 +08:00 |
|
Aurélien Geron
|
568aea61e0
|
Remove {nbsp} twice
|
2019-09-02 21:35:33 +08:00 |
|
Aurélien Geron
|
feb1dcf4a6
|
Merge pull request #26 from thierryherrmann/chap09_likelihood_fix
chap09 likelihood fix
|
2019-09-02 15:46:46 +08:00 |
|
Aurélien Geron
|
2c1c67e89a
|
Replace 1^st^ with 1st.
|
2019-09-02 15:22:35 +08:00 |
|
Thierry Herrmann
|
497ec6d2aa
|
chap09 likelihood fix
|
2019-09-01 12:55:16 -04:00 |
|
Aurélien Geron
|
15b9e4507e
|
Add a doc describing the changes in the 2nd edition
|
2019-08-27 16:18:40 +08:00 |
|
Aurélien Geron
|
0c4a9820fd
|
Update to nltk 3.4.5 (security patch)
|
2019-08-27 13:26:02 +08:00 |
|
Aurélien Geron
|
ea3d17c063
|
Fix execution_counts
|
2019-08-12 14:51:06 +08:00 |
|
Aurélien Geron
|
87ea670bef
|
Iris-Virginica => Iris virginica (& same for versicolor and setosa)
|
2019-08-12 14:45:33 +08:00 |
|
Aurélien Geron
|
aad780b74b
|
Fix Actual vs Forecast in diagrams, fixes #21
|
2019-08-03 12:11:40 +08:00 |
|
Aurélien Geron
|
281366776f
|
Add 19_training_and_deploying_at_scale.ipynb, update index.ipynb
|
2019-07-13 11:41:30 +02:00 |
|
Aurélien Geron
|
8f7a2c355b
|
Display Fashion MNIST
|
2019-07-13 11:31:17 +02:00 |
|
Aurélien Geron
|
6c0cd5d2df
|
Improve the code to forecast the next 10 time steps
|
2019-07-13 11:28:50 +02:00 |
|
Aurélien Geron
|
f8f2b9e4bb
|
Add (USD) after GDP per capita
|
2019-07-10 17:08:12 +02:00 |
|
Aurélien Geron
|
48dfa2ce67
|
tf.image.resize_image_with_crop_or_pad was renamed to tf.image.resize_with_crop_or_pad
|
2019-06-27 10:10:22 +02:00 |
|
Aurélien Geron
|
abf3bba2d5
|
Clarify DenseTranspose
|
2019-06-10 17:42:31 +08:00 |
|
Aurélien Geron
|
400920f0aa
|
SGD now defaults to lr=0.01 so use 1e-3 explicitly
|
2019-06-10 10:53:32 +08:00 |
|
Aurélien Geron
|
3db31444cd
|
SGD now defaults to lr=0.01 so use 1e-3 explicitely
|
2019-06-10 10:48:00 +08:00 |
|
Aurélien Geron
|
1973371b19
|
Use causal padding in WaveNet and fix forecast
|
2019-06-09 23:56:19 +08:00 |
|
Aurélien Geron
|
01784d2f98
|
Do not use a layer as an activation function, especially with weights
|
2019-06-09 20:08:53 +08:00 |
|
Aurélien Geron
|
0023162fc5
|
bst_n_estimators should be argmin + 1
|
2019-06-08 21:59:55 +08:00 |
|
Aurélien Geron
|
e9ae4997a5
|
Update requirements, e.g. add tf-agents, tf serving api, etc.
|
2019-06-08 12:01:58 +08:00 |
|
Aurélien Geron
|
b6306b7d48
|
Update ch. 10 code to latest TF nightly version, eg. default SGD learning rate is now 0.001
|
2019-06-08 11:43:48 +08:00 |
|
Aurélien Geron
|
0615a6f0a1
|
LayerNormalization is not experimental anymore
|
2019-06-06 11:06:13 +08:00 |
|
Aurélien Geron
|
fb47fab0c8
|
Update docker image for 2nd edition
|
2019-06-04 16:43:14 +08:00 |
|
Aurélien Geron
|
f680e49ea2
|
Replace SGD with SVC in OvA vs OvO section
|
2019-05-28 15:21:49 +08:00 |
|
Aurélien Geron
|
efa16f02ec
|
Add breakout.gif
|
2019-05-28 09:31:22 +08:00 |
|
Aurélien Geron
|
4c3b7b9b06
|
Save agent's breakout performance to an animated gif
|
2019-05-28 09:30:16 +08:00 |
|
Aurélien Geron
|
3ef350ab4c
|
Fix figure name and clarify a couple code examples
|
2019-05-27 20:35:00 +08:00 |
|
Aurélien Geron
|
c5f4b41cf5
|
Fix breakout plot
|
2019-05-26 23:56:49 +08:00 |
|
Aurélien Geron
|
2edbb6e9d4
|
Add Reinforcement Learning notebook
|
2019-05-26 23:30:39 +08:00 |
|
Aurélien Geron
|
80279c4477
|
Make notebook 09 deterministic, and avoid non-convergence warnings
|
2019-05-19 16:19:29 +07:00 |
|
Aurélien Geron
|
c702bfbcab
|
Fix ladybug.png so imageio can load it without chunk_eXIf issues
|
2019-05-19 15:03:42 +07:00 |
|
Aurélien Geron
|
73a36f335f
|
Add warning about TF issue regarding DenseFeatures and the Functional API, fixes #6
|
2019-05-15 20:23:24 +08:00 |
|
Aurélien Geron
|
ec3136f59f
|
Add index=housing.index when wrapping array in a DataFrame, fixes #426
|
2019-05-12 21:28:56 +08:00 |
|
Aurélien Geron
|
f0f83903bd
|
Fix the transformer (use final encoder outputs)
|
2019-05-10 21:30:18 +08:00 |
|
Aurélien Geron
|
7b1c890195
|
Stop using learning_phase: it's a global variable, and currently broken anyway
|
2019-05-09 16:26:38 +08:00 |
|
Aurélien Geron
|
a758ea50b0
|
Rename the figure paths from *_diagram.png to *_plot.png
|
2019-05-09 16:25:40 +08:00 |