Commits · 95833a7953dff38b6bd66e4075b5b3c0d6a5f7fd · Patrick Skillman-Lawrence / mhc_rank

Dec 19, 2017
- Improve docs · 95833a79
  Tim O'Donnell authored 7 years ago
  
  95833a79
- Better docstrings · ae3d7e2f
  Tim O'Donnell authored 7 years ago
  
  ae3d7e2f
Dec 10, 2017
- Refactor to simplify directory structure · 4ea4173a
  Tim O'Donnell authored 7 years ago
  
  4ea4173a
- formatting changes · 6f6070b2
  Tim O'Donnell authored 7 years ago
  
  6f6070b2
Dec 04, 2017
- Even faster peptide encoding · b2b032c7
  Tim O'Donnell authored 7 years ago
  
  b2b032c7
- Address code review comments · 0f9394e2
  Tim O'Donnell authored 7 years ago
  
  0f9394e2
- Update downloads to include new models · 6dfa6417
  Tim O'Donnell authored 7 years ago
  
  6dfa6417
Dec 01, 2017
- bump mhcflurry version · bdc672f4
  Tim O'Donnell authored 7 years ago
  
  pre-1.0
  
  bdc672f4
- lazy load keras · dd3e61e2
  Tim O'Donnell authored 7 years ago
  
  dd3e61e2
- Fix encoding for 0-length sequences · 4bb0738b
  Tim O'Donnell authored 7 years ago
  
  4bb0738b
Nov 30, 2017
- More speed efficient variable-length peptide encoding · 873a60a5
  Tim O'Donnell authored 7 years ago
  
  873a60a5
Nov 28, 2017
- fix · 7d323a86
  Tim O'Donnell authored 7 years ago
  
  7d323a86
- Add time remaining estimation to training script · f179489b
  Tim O'Donnell authored 7 years ago
  
  f179489b
- Free up memory: get rid of train data as models are trained · 672bf7fb
  Tim O'Donnell authored 7 years ago
  
  672bf7fb
- better progress printing · e062613c
  Tim O'Donnell authored 7 years ago
  
  e062613c
- Print progress no more often than once every 5 seconds while training · b1dbb362
  Tim O'Donnell authored 7 years ago
  
  b1dbb362
- Clearer output during training · c0fbebe1
  Tim O'Donnell authored 7 years ago
  
  c0fbebe1
- Experimental support for parallelization (Theano only) in train_allele_specific_models_command · 5b4dfbd2
  Tim O'Donnell authored 7 years ago
  
  5b4dfbd2
- implement Class1AffinityPredictor.merge() method · c6990666
  Tim O'Donnell authored 7 years ago
  
  c6990666
- arg tweak · ffcb693f
  Tim O'Donnell authored 7 years ago
  
  ffcb693f
- Add minibatch_size hyperparameter · 33b1b5d4
  Tim O'Donnell authored 7 years ago
  
  33b1b5d4
- Change prediction batch_size from 32 to 4096 for better performance · 38adfb4d
  Tim O'Donnell authored 7 years ago
  
  38adfb4d
Nov 27, 2017
- peptide encoding optimization · cbb649d6
  Tim O'Donnell authored 7 years ago
  
  cbb649d6
- Much faster performance (~5X improvement) in amino acid encoding · 41985338
  Tim O'Donnell authored 7 years ago
  
  41985338
Nov 26, 2017
- more efficient peptide encoding in calibrate_percentile_ranks · 1ca484c7
  Tim O'Donnell authored 7 years ago
  
  1ca484c7
Nov 25, 2017
- Add cross-validation GENERATE.sh and related scripts · 55e3cfc1
  Tim O'Donnell authored 7 years ago
  
  55e3cfc1
- Fix tests, add logging to calibrate_percentile_ranks, add SIGUSR1 trick to train command · ac3f874f
  Tim O'Donnell authored 7 years ago
  
  ac3f874f
- README update (mhctools mention) and percentile rank tweaks · 72362eb2
  Tim O'Donnell authored 7 years ago
  
  72362eb2
Nov 24, 2017
- fix · 2b2ce0ce
  Tim O'Donnell authored 7 years ago
  
  2b2ce0ce
- Add throw argument to percentile_ranks · e4c21bcb
  Tim O'Donnell authored 7 years ago
  
  e4c21bcb
Nov 15, 2017
- tweak · 868caf01
  Tim O'Donnell authored 7 years ago
  
  868caf01
- comment tweak · 4fbdd62c
  Tim O'Donnell authored 7 years ago
  
  4fbdd62c
- Support percentile ranks · 99e74eeb
  Tim O'Donnell authored 7 years ago
  
  * First cut of supporting quantiles ("percent ranks") per allele. Closes #87. * Some refactoring of amino_acid.py This builds on #114
  99e74eeb
Nov 14, 2017
- fix test · ed9372c2
  Tim O'Donnell authored 7 years ago
  
  ed9372c2
- Support BLOSUM62 encoding · 984120cf
  Tim O'Donnell authored 7 years ago
  
  984120cf
Nov 13, 2017
- Update README with conda instructions. Closes #110 · c1ea03b4
  Tim O'Donnell authored 7 years ago
  
  c1ea03b4
- Bump version, upgrade to Keras 2.0.9 · 3829f81e
  Tim O'Donnell authored 7 years ago
  
  * Switch from JSON to YAML for hyperparameters * Pin tensorflow (hoping will help with #110)
  3829f81e
Nov 08, 2017
- If output models directory doesn't exist, create it · 7cd4b795
  Julia K authored 7 years ago
  
  7cd4b795
- Fixing code typo: allelle -> allele. · c4f42dd3
  Julia K authored 7 years ago
  
  c4f42dd3
Oct 13, 2017
- Remove antigen presentation submodule (for now) · 1d37637c
  Tim O'Donnell authored 7 years ago
  
  1d37637c