Commits · 474383af63a6d6b49428c230f23f2f973884d5f4 · Patrick Skillman-Lawrence / mhc_rank

Jun 02, 2016
- default params -- do they work for training quickly? · 474383af
  Alex Rubinsteyn authored 8 years ago
  
  474383af
- mysterious xc3 non-ASCII character possibly vanquished · ab08abf7
  Alex Rubinsteyn authored 8 years ago
  
  ab08abf7
- Moving shared functionality from deep within Dataset to helper functions in common.py · 9508c208
  Alex Rubinsteyn authored 8 years ago
  
  9508c208
- don't use batch normalization in small dataset unit tests · 9896944c
  Alex Rubinsteyn authored 8 years ago
  
  9896944c
- pass dropout to Embedding instead of creating Dropout layer · 53496d26
  Alex Rubinsteyn authored 8 years ago
  
  53496d26
- Merge pull request #27 from hammerlab/add-batch-normalization · 837d7b54
  Alex Rubinsteyn authored 8 years ago
  
  added batch normalization between layers, changed default weight initialization
  837d7b54
Jun 01, 2016

added batch normalization between layers and changed default initialization to... · 3bc225c0

Alex Rubinsteyn authored 8 years ago

added batch normalization between layers and changed default initialization to divide by sum of fan_in + fan_out instead of just fan_in

3bc225c0

May 21, 2016
- updates to dataset size sensitivity script · 52a88ace
  Alex Rubinsteyn authored 8 years ago
  
  52a88ace
- run script through Python 3 · f0fdbc98
  Alex Rubinsteyn authored 8 years ago
  
  f0fdbc98
- lots of tweaks to dataset size sensitivity script · 328d2c12
  Alex Rubinsteyn authored 8 years ago
  
  328d2c12
May 19, 2016
- using stratified sampling for dataset size sensitivity · d122c743
  Alex Rubinsteyn authored 8 years ago
  
  d122c743
- use correct ScoreSet · 91a7ed7d
  Alex Rubinsteyn authored 8 years ago
  
  91a7ed7d
- moved score_set back to experiments since not using it · 499420cc
  Alex Rubinsteyn authored 8 years ago
  
  499420cc
- dont' add --max-ic50 to args twice · 95de1b69
  Alex Rubinsteyn authored 8 years ago
  
  95de1b69
- using mhcflurry.args in dataset-size-sensitivity · f7039fb4
  Alex Rubinsteyn authored 8 years ago
  
  f7039fb4
- added random shrinkage for right censored affinities · 5e820f66
  Alex Rubinsteyn authored 8 years ago
  
  5e820f66
- faster algorithm for converting Dataset to nested dictionary · 8c5a4ca5
  Alex Rubinsteyn authored 8 years ago
  
  8c5a4ca5
May 18, 2016
- added intersection operator to Dataset · 180575f6
  Alex Rubinsteyn authored 8 years ago
  
  180575f6
May 17, 2016
- added cross_validation iterator to Dataset · 4b337b22
  Alex Rubinsteyn authored 8 years ago
  
  4b337b22
- added helpers to Dataset · 37e0372b
  Alex Rubinsteyn authored 8 years ago
  
  37e0372b
May 14, 2016
- don't rely on Params.__dict__ · 361dc375
  Alex Rubinsteyn authored 8 years ago
  
  361dc375
- weirdly some versions of Python don't give __dict__ properties to namedtuples · b42fbf1d
  Alex Rubinsteyn authored 8 years ago
  
  b42fbf1d
- moved Params definition down · e766e86a
  Alex Rubinsteyn authored 8 years ago
  
  e766e86a
- added Params namedtuple, helper to create all combinations of hyperparams, and simple test · f13a6c26
  Alex Rubinsteyn authored 8 years ago
  
  f13a6c26
- added Dataset tests · 2c380b44
  Alex Rubinsteyn authored 8 years ago
  
  2c380b44
May 13, 2016
- Merge pull request #26 from hammerlab/dataset · c894b11f
  Alex Rubinsteyn authored 8 years ago
  
  Generalizing AlleleData into Dataset
  c894b11f
- added min_ arguments to to_dense_pMHC_matrix · a2e339d5
  Alex Rubinsteyn authored 8 years ago
  
  a2e339d5
- use Dataset.from_csv in matrix-completion-accuracy · ca7f46de
  Alex Rubinsteyn authored 8 years ago
  
  ca7f46de
- got rid of matrix completion helpers · a20c6ee0
  Alex Rubinsteyn authored 8 years ago
  
  a20c6ee0
- fixed import path for indices_to_hotshot_encoding · 70588389
  Alex Rubinsteyn authored 8 years ago
  
  70588389
- removing all references to mhcflurry.data from experiments scripts · e5a27666
  Alex Rubinsteyn authored 8 years ago
  
  e5a27666
- always return parser from args helpers · c6e1410f
  Alex Rubinsteyn authored 8 years ago
  
  c6e1410f
- added random negative samples to training · ce1bbda1
  Alex Rubinsteyn authored 8 years ago
  
  ce1bbda1
- added dataset-size-sensitivity script · 7fef98d0
  Alex Rubinsteyn authored 8 years ago
  
  7fef98d0
May 11, 2016
- small fixes · 629baf60
  Alex Rubinsteyn authored 8 years ago
  
  629baf60
May 10, 2016
- working on liberating the titration script from the experiments directory · d1aa3d76
  Alex Rubinsteyn authored 8 years ago
  
  d1aa3d76
- fixed obscure shape bug on empty arrays · fa353a34
  Alex Rubinsteyn authored 8 years ago
  
  fa353a34
May 09, 2016
- fixing training script · 345869a0
  Alex Rubinsteyn authored 8 years ago
  
  345869a0
- added typechecks as requirement · 4f0f0820
  Alex Rubinsteyn authored 8 years ago
  
  4f0f0820
- only train for 200 epochs in unit test · 8a19a894
  Alex Rubinsteyn authored 8 years ago
  
  8a19a894