anton/mlpractical
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Commit Graph

  • 822b946033 Adding MSD data providers. Matt Graham 2017-01-21 19:34:01 +0000
  • 510a1ca7be Adding CIFAR data providers. Matt Graham 2017-01-19 18:51:18 +0000
  • c4a662465e Clarifying initializers description: notebook save failed in previous commit. Matt Graham 2017-01-19 10:49:10 +0000
  • 223bd971d1 Clarifying initializers description. Matt Graham 2017-01-18 13:14:03 +0000
  • 23e015085f Adding optional contrib modules exercise. Matt Graham 2017-01-17 18:26:33 +0000
  • e0262e43fc Adding TensorFlow install instructions: attempt 2. Matt Graham 2017-01-17 12:57:16 +0000
  • dfd7414791 Merge branch 'mlp2016-7/master' of https://github.com/CSTR-Edinburgh/mlpractical into mlp2016-7/master Matt Graham 2017-01-17 12:55:28 +0000
  • 8b756fce5d Adding TensorFlow install instructions. Matt Graham 2017-01-17 12:55:00 +0000
  • 8390e4a822 Updating Autoencoders notebook with solutions. Matt Graham 2016-12-26 18:58:30 +0000
  • ca9f1e6424 Clearing outputs from TensorFlow intro notebook. Matt Graham 2016-12-22 23:52:30 +0000
  • ae666f895a Adding figures for TensorFlow intro notebook. Matt Graham 2016-12-22 23:42:21 +0000
  • 8c08b0d5ea Adding TensorFlow introduction notebook. Matt Graham 2016-12-22 23:40:46 +0000
  • 6892424a43 Fix __repr__ implementation for SingleLayerModel (#46) Jonathan Jouty 2016-11-22 22:21:45 +0000
  • d41063dca6 Rerunning dropout experiment with non-stochastic evaluation. Matt Graham 2016-11-09 19:58:26 +0000
  • 77a96cfc01 Fixing use of global rng in DropoutLayer fprop. Matt Graham 2016-11-09 19:55:11 +0000
  • 9c86d00078 Changing default optimiser behaviour to use non-stochastic evaluation for stochastic layers. Matt Graham 2016-11-09 19:50:32 +0000
  • d290f4513d Fixing inconsistent channel ordering in constructor and docstring shape descriptions. Matt Graham 2016-11-09 19:40:46 +0000
  • 60edabd261 Fixing missing bias term from model visualisation. Matt Graham 2016-11-05 22:21:44 +0000
  • 905545fd23 Adding notebook for 7th lab on autoencoders. Matt Graham 2016-11-05 22:20:29 +0000
  • 9bb4d7560b Adding dropout and maxout layers to module. Matt Graham 2016-11-05 22:19:47 +0000
  • fa9ef33460 Adding dropout and maxout notebook answers. Matt Graham 2016-11-05 22:19:28 +0000
  • 02e29ddcc6 Adding time-dependent learning rule schedulers. Matt Graham 2016-11-03 18:31:59 +0000
  • 2b20484649 Adding adaptive learning rules. Matt Graham 2016-11-03 18:31:37 +0000
  • 6f4aa5fd5f Adding regularisation penalties module. Matt Graham 2016-11-03 18:28:57 +0000
  • 057a25ed07 Adding autoencoder wrapper data provider. Matt Graham 2016-11-03 18:28:41 +0000
  • 2a4364d226 Adding necessary imports. Matt Graham 2016-11-02 03:37:22 +0000
  • 534c4744ca Fixing incorrect use of module relative class reference. Matt Graham 2016-11-02 03:36:53 +0000
  • 891b942890 Merge branch 'mlp2016-7/master' of https://github.com/CSTR-Edinburgh/mlpractical into mlp2016-7/master Matt Graham 2016-11-02 02:55:37 +0000
  • a575efb0ee Adding instructions on running notebooks remotely. Matt Graham 2016-11-02 02:55:10 +0000
  • e0114b9aab Adding notebook with convolutional layer tests. Matt Graham 2016-11-02 01:14:27 +0000
  • da9ee58609 Adding reshape layer. Matt Graham 2016-11-02 01:13:22 +0000
  • d5ae8b2caf Merge pull request #43 from jfainberg/patch-1 Matt Graham 2016-10-29 21:17:24 +0100
  • bea3502210 Minor typos in momentum docstring Joachim 2016-10-29 15:57:00 +0100
  • a6f775b1ff Adding notebook for lab 6. Matt Graham 2016-10-27 21:43:35 +0100
  • 1efe9de847 Adding code for lab 5 notebook answers. Matt Graham 2016-10-27 21:43:08 +0100
  • 95d4028bb7 Adding stochastic layer functionality in preparation for dropout implementation. Matt Graham 2016-10-27 21:36:57 +0100
  • 7e1250a82b Adding bash script to secure Jupyter notebook server on DICE. Matt Graham 2016-10-27 21:35:55 +0100
  • dfcce94242 Clarifying docstring in scheduler. Matt Graham 2016-10-27 11:24:00 +0100
  • ae999b43e4 Fixing spacing typo. Matt Graham 2016-10-21 01:37:00 +0100
  • 9a182d4be1 Correcting regularisation coefficient values to test. Matt Graham 2016-10-21 01:35:50 +0100
  • cf6417f917 Removing old transfer function notebook. Matt Graham 2016-10-21 22:06:37 +0100
  • 586c7477eb Removing old regularisation notebook. Matt Graham 2016-10-21 01:25:08 +0100
  • 167508cc63 Adding new lab 5 notebook. Matt Graham 2016-10-21 01:24:37 +0100
  • ab060d556c Switching to parameter penalty rather than cost naming. Matt Graham 2016-10-21 01:24:04 +0100
  • a353526790 Adding augmented MNIST data provider. Matt Graham 2016-10-20 13:26:09 +0100
  • a021ce585b Applying softmax stability fix to cost gradient. Matt Graham 2016-10-20 13:23:19 +0100
  • b4b00e0c65 Using more stable softmax implementation by subtracting max. Matt Graham 2016-10-18 13:54:39 +0100
  • 870c04f664 Modifying optmiser.train call in notebook to use new interface with run_time. Matt Graham 2016-10-18 13:45:28 +0100
  • 0f496306c2 Fixing RadialBasisFunctionLayer __repr__ bug. Matt Graham 2016-10-18 13:45:05 +0100
  • 237cff12db Making sure input_dims set converted to list. Matt Graham 2016-10-18 13:44:30 +0100
  • 5d4c979ef3 Changing code blocks to use small p Python. Matt Graham 2016-10-17 23:36:21 +0100
  • f13ba91f82 Dealing with case where number of epochs less than stats interval. Matt Graham 2016-10-17 23:28:01 +0100
  • 0cec9920f3 Adding coursework files to master branch. Matt Graham 2016-10-14 03:56:06 +0100
  • 4863599bed Renaming old notebooks to reflect new numbering. Matt Graham 2016-10-14 03:40:43 +0100
  • 0b9898d72c Adding new fourth lab notebook. Matt Graham 2016-10-14 03:39:19 +0100
  • 43ad2a00ba Adding RBF layer definition for 4th lab. Matt Graham 2016-10-14 03:36:46 +0100
  • 2e5a1a4c4f Renaming MNIST eval set to test set for consistency with slides. Matt Graham 2016-10-14 03:36:16 +0100
  • 4c4cf2cf27 Bug fix: incorrect use of zip in MomentumLearningRule. Matt Graham 2016-10-12 23:11:38 +0100
  • a59d37dc69 Fixing softmax index typo. Matt Graham 2016-10-11 15:39:35 +0100
  • cafed216d7 Fixing typo in description of back-propagation. Matt Graham 2016-10-11 13:40:29 +0100
  • a6fcdc2894 Fixing typo in initialiser assert. Matt Graham 2016-10-10 10:09:29 +0100
  • cfad36c91f Removing final backpropagation of gradients to inputs as unnecessary. Matt Graham 2016-10-10 09:27:02 +0100
  • a24ff90f71 Adding Nesterov momentum learning rule. Matt Graham 2016-10-10 09:25:52 +0100
  • 259b000cba Adding schedulers to optimiser. Matt Graham 2016-10-10 09:25:33 +0100
  • 18f893d864 Adding Glorot and Bengio parameter initialiser. Matt Graham 2016-10-10 09:24:32 +0100
  • 302d44371a Adding new schedulers module. Matt Graham 2016-10-10 09:23:53 +0100
  • fa2f6b195c Correcting docstring typo. Matt Graham 2016-10-08 22:19:49 +0100
  • 1fd2b84353 Renaming third lab notebook. Matt Graham 2016-10-07 07:01:26 +0100
  • ee0c6ce1d5 Updating third lab notebook on multilayer models. Matt Graham 2016-10-07 07:00:47 +0100
  • b1a031deac Adding solutions for second lab notebook. Matt Graham 2016-10-07 07:00:10 +0100
  • 9863c75b05 Adding back propagation block diagram figure and source. Matt Graham 2016-10-07 06:24:18 +0100
  • 25263e1391 Including initial pre-training statistics. Matt Graham 2016-10-07 06:23:15 +0100
  • e05717e3d0 Adding SoftmaxLayer definition. Matt Graham 2016-10-07 06:22:12 +0100
  • 8d6f37668f Adding ability to adjust batch size after initialisation and resetting to initial state (order). Matt Graham 2016-10-07 06:21:34 +0100
  • f979aae573 Adding lab start up instructions. Matt Graham 2016-10-04 12:59:06 +0100
  • e8333c02f1 Squashed commit of fixes to allow lab 1 notebook to render on Github. Matt Graham 2016-10-03 12:20:24 +0100
  • baafac723e Attempt to make updated lab 1 notebook display on Github. Matt Graham 2016-10-03 11:49:41 +0100
  • a6c2eb2ae1 Adding first lab notebook solutions. Matt Graham 2016-10-03 11:36:33 +0100
  • 57ae165135 Removing previous second lab notebook. Matt Graham 2016-09-30 02:54:32 +0100
  • 881b2e7aee Finishing material for second lab notebook. Matt Graham 2016-09-30 02:53:54 +0100
  • 4ef1428447 Switching from 'cost' to 'error' for consistency with slides. Matt Graham 2016-09-30 02:53:13 +0100
  • f1ed11a325 Initial updates to lab 2 notebook. Matt Graham 2016-09-28 05:07:40 +0100
  • 2f9253bd21 Adding CCPP data and data provider. Matt Graham 2016-09-28 05:07:01 +0100
  • 15425cff0e Further clarifying reload vs. import note. Matt Graham 2016-09-27 19:14:44 +0100
  • 09901921ff Further clarifying reload vs. import note. Matt Graham 2016-09-27 19:14:44 +0100
  • a0c2f0db1b Changing to using remove rather than clean in quota issue fix based on lab feedback. Matt Graham 2016-09-27 18:46:59 +0100
  • 8a46ac7fc0 Adding clarifications to set up instructions based on feedback from lab. Matt Graham 2016-09-27 18:46:35 +0100
  • 137782a06a Adding note about quota issues. Matt Graham 2016-09-27 10:26:13 +0100
  • b86961d09e Fix to make sure MLP_DATA_DIR defined in current session. Matt Graham 2016-09-26 12:20:57 +0100
  • f2b6243a84 Adding Python/NumPy tutorial credits. Matt Graham 2016-09-23 13:13:55 +0100
  • 7fd23d81cf Adding link to Stanford numpy tutorial and readability improvements. Matt Graham 2016-09-23 13:08:16 +0100
  • 1fae6cb8a2 Fixes to typos and some minor readability improvments. Matt Graham 2016-09-23 12:47:33 +0100
  • 4b525bbadb Changing to using remove rather than clean in quota issue fix based on lab feedback. Matt Graham 2016-09-27 18:46:59 +0100
  • 910e3b7a0b Adding clarifications to set up instructions based on feedback from lab. Matt Graham 2016-09-27 18:46:35 +0100
  • 762498dd17 Adding note about quota issues. Matt Graham 2016-09-27 10:26:13 +0100
  • 85d692c974 Fix to make sure MLP_DATA_DIR defined in current session. Matt Graham 2016-09-26 12:20:57 +0100
  • ec405f8fd9 Adding Python/NumPy tutorial credits. Matt Graham 2016-09-23 13:13:55 +0100
  • 06f1ba6906 Adding link to Stanford numpy tutorial and readability improvements. Matt Graham 2016-09-23 13:08:16 +0100
  • 64275866ef Fixes to typos and some minor readability improvments. Matt Graham 2016-09-23 12:47:33 +0100
  • 0b9c88ff20 Merge branch 'mlp2016-7/lab1' of https://github.com/CSTR-Edinburgh/mlpractical into mlp2016-7/lab1 Matt Graham 2016-09-22 16:52:27 +0100