754 Commits

Author SHA1 Message Date
Aurélien Geron
cfd0837f5c Add LunarLander-v2 Policy Gradients exercise solution 2021-03-20 10:04:52 +13:00
Aurélien Geron
8cf72920b9 Upgrade Pillow library 2021-03-19 10:56:39 +13:00
Aurélien Geron
9d33d65ef0 logs["loss"] is the mean loss, not the batch loss anymore (since TF 2.2), fixes #188 2021-03-19 10:50:13 +13:00
Aurélien Geron
e9b5dce122 Fix auto-fire, add exercises, explain Space Invaders delta 2021-03-18 22:16:38 +13:00
Aurélien Geron
86702573c6 Remove old unused code, and clarify semantic hashing section, fixes #246 2021-03-18 15:39:31 +13:00
Aurélien Geron
2f2f9c8dd3 Save positional_embedding_plot figure using save_fig() 2021-03-11 15:11:44 +13:00
Aurélien Geron
b46db0e1a1 To get GPU acceleration with GRUs, do not use recurrent_dropout 2021-03-11 15:07:23 +13:00
Aurélien Geron
d423e2254e Divide MSE_node by m_node in equation 6-4 2021-03-10 18:45:05 +13:00
Aurélien Geron
c98ee19363 Fix AtariPreprocessingWithAutoFire typo 2021-03-10 10:45:24 +13:00
Aurélien Geron
e4bf060ea5 Merge branch 'master' of github.com:ageron/handson-ml2 2021-03-10 09:17:50 +13:00
Aurélien Geron
7719ff276a Replace x0 with x'0 in right side of figure 5-2 2021-03-10 09:17:15 +13:00
Aurélien Geron
4272b5313a Merge pull request #401 from gae7/ch_12
Update 12_custom_models_and_training_with_tensorflow.ipynb
2021-03-10 08:45:12 +13:00
Aurélien Geron
b2d197711f Change np.fromfunction() example and fix typo in explanation, fixes #403 2021-03-10 08:30:14 +13:00
Aurélien Geron
f25c66cc3a Merge branch 'master' of github.com:ageron/handson-ml2 2021-03-09 22:21:21 +13:00
Aurélien Geron
dd94101c5d Speed up training: I tuned learning rate for DQN variants, and added auto-FIRE for Blockout. Fixes #117 2021-03-09 22:21:08 +13:00
Aurélien Geron
b3b1ed0a57 Merge pull request #402 from ll7/patch-1
update environment gpu support
2021-03-09 21:19:13 +13:00
ll7
7024fabc46 update environment gpu support
The docker-compose was described with an old envirnoment file?
2021-03-08 16:36:42 +01:00
gae7
dc584421f6 Update 12_custom_models_and_training_with_tensorflow.ipynb 2021-03-07 11:25:47 +01:00
Aurélien Geron
80eec21242 Update the Docker image compressed sizes (shrunk by about 500MB each) 2021-03-07 10:35:58 +13:00
Aurélien Geron
0c17db555c Add a comment about apt clean 2021-03-06 20:35:57 +13:00
Aurélien Geron
9056707a18 Shrink the docker image size 2021-03-06 20:33:53 +13:00
Aurélien Geron
d399affb5c Shrink the docker image size 2021-03-06 19:51:19 +13:00
Aurélien Geron
5da46e924c Add link to contributors 2021-03-05 23:27:05 +13:00
Aurélien Geron
15da7b4cee Update thanks 2021-03-05 23:25:20 +13:00
Aurélien Geron
7bcdff0d8f Update thanks 2021-03-05 23:22:49 +13:00
Aurélien Geron
b59bd8c745 Simplify and organize the instructions 2021-03-05 23:03:23 +13:00
Aurélien Geron
931c2c9558 Clarify the Docker documentation, in particular for GPU support 2021-03-05 22:37:37 +13:00
Aurélien Geron
1de2af7739 Remove the need to tag after pull 2021-03-05 18:49:37 +13:00
Aurélien Geron
2c4b4ad0dd Merge branch 'master' of github.com:ageron/handson-ml2 2021-03-05 18:47:53 +13:00
Aurélien Geron
05d2c1ea9b Remove the need to tag after pull 2021-03-05 18:47:37 +13:00
Aurélien Geron
e96e10c7f2 Add info to pull GPU image from Docker Hub 2021-03-05 17:33:59 +13:00
Aurélien Geron
e10a807907 Set max_features="sqrt" in Decision Tree rather than BaggingClassifier 2021-03-04 23:06:21 +13:00
Aurélien Geron
33f9ff10b4 Use default splitter="best" instead of splitter="random", fixes #340 2021-03-04 15:17:19 +13:00
Aurélien Geron
e2edfb274b gradients(dfdx, [x, y]) instead of gradients(dfdx, [3., 4.]), fixes #399 2021-03-04 09:11:09 +13:00
Aurélien Geron
d6c2ba1b8f Add Dockerfile.gpu for GPU support 2021-03-03 22:19:27 +13:00
Aurélien Geron
abd75d7fd7 Explain how to use the docker image from Docker Hub
Also make it clearer that `make` is optional.
2021-03-03 11:15:23 +13:00
Aurélien Geron
99361d1f82 Merge branch 'master' of github.com:ageron/handson-ml2 2021-03-02 22:47:23 +13:00
Aurélien Geron
9abeffd21c Remove work_in_progress directory 2021-03-02 22:47:09 +13:00
Aurélien Geron
f65e2e7ad9 Clarify that 'make' is optional 2021-03-02 22:27:20 +13:00
Aurélien Geron
55ee303e56 Merge pull request #275 from ibeauregard/changes-chap11
(Chapter 11) Adjust computation of steps per epoch
2021-03-02 22:12:35 +13:00
Aurélien Geron
46307aced0 Merge branch 'master' of github.com:ageron/handson-ml2 2021-03-02 18:14:59 +13:00
Aurélien Geron
9286e4f04f Fix comment and simplify 'neighbors' code 2021-03-02 18:14:12 +13:00
Aurélien Geron
e506421ff2 Merge pull request #249 from AndC314/patch-1
Update 01_the_machine_learning_landscape.ipynb
2021-03-02 17:51:00 +13:00
Aurélien Geron
6f1ede95a2 Merge pull request #245 from baseplate77/master
Update 16_nlp_with_rnns_and_attention.ipynb
2021-03-02 17:50:31 +13:00
Aurélien Geron
63a9c086b6 Merge pull request #237 from ajrator/patch-1
Modify import of urllib to urllib.request
2021-03-02 17:49:51 +13:00
Aurélien Geron
af6064de0c Merge pull request #232 from ibeauregard/changes-chap5
Use object attribute instead of global constant
2021-03-02 17:49:17 +13:00
Aurélien Geron
0a40b6b051 Merge pull request #225 from ngaro/dockerrebuild
Change rebuild target
2021-03-02 17:48:37 +13:00
Aurélien Geron
b2a1f56988 Update chapter 17's name 2021-03-02 17:47:26 +13:00
Aurélien Geron
4af17a6aac Merge pull request #222 from guoci/patch-1
Update index.ipynb
2021-03-02 17:45:55 +13:00
Aurélien Geron
4963911f4f Merge pull request #229 from ibeauregard/changes-chap4
Remove useless code
2021-03-02 17:27:54 +13:00