 René Knaebel
|
5741f8ee0e
|
fix staggered training
|
2017-10-06 10:38:00 +02:00 |
|
René Knaebel
|
508667d1d0
|
add server classification model
|
2017-10-05 15:26:53 +02:00 |
|
René Knaebel
|
345afbaef5
|
add error bar vis, comment unused parameters from parser
|
2017-10-05 14:50:59 +02:00 |
|
René Knaebel
|
b24fa770f9
|
change hyperband to count minimal val_loss over all losses
|
2017-10-05 12:55:46 +02:00 |
|
René Knaebel
|
371a1dad05
|
add hyperband savefile config, minor change of parameter name
|
2017-10-03 18:58:54 +02:00 |
|
René Knaebel
|
f02e0b7f99
|
fix hyperband wrong variable names
|
2017-09-29 23:34:39 +02:00 |
|
René Knaebel
|
a1e553f8f1
|
fix rerun wrong path
|
2017-09-29 23:03:07 +02:00 |
|
René Knaebel
|
605447440f
|
refactor hyperband implementation
|
2017-09-29 22:59:57 +02:00 |
|
René Knaebel
|
090c89a127
|
add retrain mode
|
2017-09-28 12:23:22 +02:00 |
|
René Knaebel
|
b157ca6a19
|
add first version of model averaging visualization
|
2017-09-26 19:25:37 +02:00 |
|
René Knaebel
|
607d74998c
|
add tsne (does not work with big data)
fix model loading with custom selu function
|
2017-09-22 10:01:12 +02:00 |
|
René Knaebel
|
f2845e635e
|
fix validation split removal by loading h5data into memory
|
2017-09-17 09:56:18 +02:00 |
|
René Knaebel
|
ec5a1101be
|
remove model selection based on validation loss
|
2017-09-16 15:25:34 +02:00 |
|
René Knaebel
|
7f49021a63
|
refactor training - separate staggered training; make differences as small as possible
|
2017-09-12 08:36:23 +02:00 |
|
René Knaebel
|
3a44efa775
|
minor changes for pauls test
|
2017-09-11 12:59:13 +02:00 |
|
René Knaebel
|
6121eac448
|
remove regularizer for conv and domain
|
2017-09-10 23:40:14 +02:00 |
|
René Knaebel
|
1cf62423e1
|
add regularization to small networks, fix model name in args, fix visualizations
|
2017-09-10 18:06:40 +02:00 |
|
René Knaebel
|
6fef2b8b84
|
refactor all visualization for pauls changes - evaluate on max windows per users
|
2017-09-08 22:59:55 +02:00 |
|
René Knaebel
|
9a51b6ea34
|
refactor test function working on full unfiltered data
|
2017-09-08 19:10:23 +02:00 |
|
René Knaebel
|
edc75f4f44
|
refactor dataset creation, split up functions
|
2017-09-08 17:11:13 +02:00 |
|
René Knaebel
|
70d00efb01
|
refactor using joblib for test results, make h5py store/load more flexible
|
2017-09-08 13:55:13 +02:00 |
|
René Knaebel
|
1ab0108c78
|
add window to file names for visualization
|
2017-09-07 17:38:21 +02:00 |
|
René Knaebel
|
595c2ea894
|
change argument interface
- add more properties for network specification
- change names for consistency
|
2017-09-07 15:53:58 +02:00 |
|
René Knaebel
|
71f218888d
|
set server to be not trainable too; refactor visualization script
|
2017-09-07 15:31:04 +02:00 |
|
René Knaebel
|
5bd8e41711
|
add staggered model training for intermediate sever prediction; refactor model return values
|
2017-09-07 14:24:55 +02:00 |
|
René Knaebel
|
2080444fb7
|
add parser argument for naming in multi model modes, minor fixes, re-set fix vals for network - need to make them flexible
|
2017-09-05 17:40:57 +02:00 |
|
René Knaebel
|
1da31cc97c
|
visualize per user stats
|
2017-09-04 13:37:26 +02:00 |
|
René Knaebel
|
3f6779fa3d
|
load names with data for per-user evaluation
|
2017-09-02 16:02:48 +02:00 |
|
René Knaebel
|
dc9180da10
|
refactor visualization, change arguments for model type and its depth
|
2017-09-01 10:42:26 +02:00 |
|
René Knaebel
|
954dfcf9f9
|
fix run scripts
|
2017-08-05 09:40:40 +02:00 |
|
René Knaebel
|
5a02f582cd
|
change model - add dense before server output in new model
add some new run scripts
|
2017-08-05 09:33:07 +02:00 |
|
René Knaebel
|
6e7dc1297c
|
fix lazy domain loading and generation process
|
2017-08-03 12:27:17 +02:00 |
|
René Knaebel
|
7f1d13658f
|
store domain embeddings while test main
|
2017-08-03 09:08:24 +02:00 |
|
René Knaebel
|
452f9e0456
|
fix array() to zeros(0)
|
2017-08-03 08:00:20 +02:00 |
|
René Knaebel
|
787f43b328
|
fix test predictions depending on model output specification
|
2017-08-03 07:51:58 +02:00 |
|
René Knaebel
|
1e781d5491
|
add argument for model outputs
BUG: need to check --new_model --model_output server
|
2017-08-02 12:58:09 +02:00 |
|
René Knaebel
|
f4da147688
|
refactor cmd argument to have single value for mode
|
2017-07-30 15:49:37 +02:00 |
|
René Knaebel
|
e24f596f40
|
add argument for using the new model architecture
|
2017-07-30 14:07:39 +02:00 |
|
René Knaebel
|
ebaeb6b96e
|
move vocab_size into implementation (not user dependent)
|
2017-07-30 13:47:11 +02:00 |
|
René Knaebel
|
d97785f646
|
replace softmax by sigmoid in final layer, also adjust dataset for that
|
2017-07-30 12:50:26 +02:00 |
|
René Knaebel
|
b0da2de0ea
|
move utils functions to new file
|
2017-07-29 19:47:02 +02:00 |
|
René Knaebel
|
820a5d1a4d
|
add new network architecture - server label moves to the middle
|
2017-07-29 19:42:36 +02:00 |
|
René Knaebel
|
18b60e1754
|
add extended test mode for embeddings
|
2017-07-17 19:30:56 +02:00 |
|
René Knaebel
|
844494eca9
|
add multi-threading for pre-processing
|
2017-07-16 09:42:52 +02:00 |
|
René Knaebel
|
336be37032
|
fix network props, add PCA to visualize main
|
2017-07-14 21:01:08 +02:00 |
|
René Knaebel
|
6b787792db
|
add custom class weights based on sklearn balance
|
2017-07-14 15:57:52 +02:00 |
|
René Knaebel
|
b35f23e518
|
add visualization for training curves, pr, roc
|
2017-07-14 14:58:17 +02:00 |
|
René Knaebel
|
d0418b9efa
|
fix hyperband adding h5data and remove param sample
|
2017-07-12 15:18:45 +02:00 |
|
René Knaebel
|
2afaccc84b
|
refactor argparser into separate file, add logger
|
2017-07-12 10:25:55 +02:00 |
|
René Knaebel
|
9f0bae33d5
|
refactor dataset generation, add callbacks
|
2017-07-11 21:06:58 +02:00 |