Commits · 237cecc971c068a8e76d4d6314650cc1d6442254 · Jon Lachmann / sgd · GitLab

Sep 05, 2015
- make bigmemory optional; fix #82 · 237cecc9
  Dustin Tran authored Sep 05, 2015
  
  237cecc9
Sep 03, 2015
- add warning message if reach max iters · ad886192
  Dustin Tran authored Sep 02, 2015
  
  ad886192
- relax input requirements, making glm compatible · d0183a13
  Dustin Tran authored Aug 23, 2015
  
  d0183a13
Aug 21, 2015
- Code for M-estimation. · e4766c2d
  ptoulis authored Aug 21, 2015
  
  e4766c2d
- Code for M-estimation. · 391275c8
  ptoulis authored Aug 21, 2015
  
  391275c8
Aug 18, 2015
- tease out generic methods as separate files · 7ce7db38
  Dustin Tran authored Aug 18, 2015
  
  7ce7db38
- clean up TODO comments · 75b8857b
  Dustin Tran authored Aug 18, 2015
  
  75b8857b
- move all library includes into basedef; clean up includes · d5b8bc75
  Dustin Tran authored Aug 18, 2015
  
  d5b8bc75
- add runtime plot · d02f430e
  Dustin Tran authored Aug 18, 2015
  
  d02f430e
Aug 16, 2015
- add predict method; make glm have compatible output as oall other built-in models · 2571a4c1
  Dustin Tran authored Aug 16, 2015
  
  2571a4c1
- add working plot_mse; add model name to sgd output · 51f17690
  Dustin Tran authored Aug 16, 2015
  
  51f17690
Aug 15, 2015
- add cox model · beee5eb5
  Dustin Tran authored Aug 14, 2015
  
  beee5eb5
Aug 13, 2015
- allow data_point class to display its index to the data set (for cox model) · 6bff0825
  Dustin Tran authored Aug 13, 2015
  
  6bff0825
Aug 12, 2015
- remove verbose comments · adefe592
  Dustin Tran authored Aug 12, 2015
  
  adefe592
- swap order of learning_rate args to follow convention of index first · 0bc343b0
  Dustin Tran authored Aug 12, 2015
  
  0bc343b0
- allow gradient to take in data index, not data point (relevant for cox model) · e98c14fa
  Dustin Tran authored Aug 12, 2015
  
  e98c14fa
- remove #include data_point header in learning rates · cea93717
  Dustin Tran authored Aug 12, 2015
  
  cea93717
Aug 11, 2015
- move delta private var to implicit_sgd derived class only · 0fcf8896
  Dustin Tran authored Aug 11, 2015
  
  0fcf8896
- remove need for grad_func(); let gradient potentially use all data · a3ac85db
  Dustin Tran authored Aug 11, 2015
  
  a3ac85db
Aug 10, 2015
- remove verbose comments · aaadd7b2
  Dustin Tran authored Aug 10, 2015
  
  aaadd7b2
Jul 31, 2015
- remove indexing into convergence list element · 6381f48c
  Dustin Tran authored Jul 11, 2015
  
  6381f48c
Jul 06, 2015
- always output converged attribute · dc7546af
  Dustin Tran authored Jul 02, 2015
  
  dc7546af
Jul 01, 2015
- implement sgd.function; improve documentation · 3dae987f
  Dustin Tran authored Jul 01, 2015
  
  3dae987f
Jun 30, 2015
- update README.md · eac49a5d
  Dustin Tran authored Jun 30, 2015
  
  eac49a5d
- update description of package · c44ead7f
  Dustin Tran authored Jun 30, 2015
  
  c44ead7f
- minor change in documentaiton of start argument; fix #58 · edce7abb
  Dustin Tran authored Jun 30, 2015
  
  edce7abb
- update README.md · 84b357be
  Dustin Tran authored Jun 30, 2015
  
  84b357be
- add classical momentum as new sgd method; #59 · a65fc476
  Dustin Tran authored Jun 30, 2015
  
  a65fc476
- add nesterov momentum as new sgd method; #59 · dacbb3c8
  Dustin Tran authored Jun 30, 2015
  
  dacbb3c8
- add demo for estimating equations · 70e14678
  Dustin Tran authored Jun 30, 2015
  
  70e14678
- fix class printing for ee model · 6ad8b1dd
  Dustin Tran authored Jun 30, 2015
  
  6ad8b1dd
- update README.md · f7262569
  Dustin Tran authored Jun 30, 2015
  
  f7262569
- split number of params from number of features; change ee_model constructor · 66c9717a
  Dustin Tran authored Jun 30, 2015
  
  66c9717a
Jun 29, 2015
- remove need for model.attrs in model.control · b482a131
  Dustin Tran authored Jun 29, 2015
  
  b482a131
- set ddim learning rate to store only a vector, not a matrix · 5911b13f
  Dustin Tran authored Jun 29, 2015
  
  5911b13f
- complete documentation for sgd output · 5ce3a084
  Dustin Tran authored Jun 29, 2015
  
  5ce3a084
- allow size to be argument in sgd_control · 0a260b70
  Dustin Tran authored Jun 29, 2015
  
  0a260b70
- fix poisson regression demo since offset no longer implemented · a15eea29
  Dustin Tran authored Jun 29, 2015
  
  a15eea29
- add explicit_sgd and implicit_sgd as subclasses of a base_sgd · bf1c96d6
  Dustin Tran authored Jun 29, 2015
  
  bf1c96d6
- split experiment argument into model_control and sgd_control · 6bc3c629
  Dustin Tran authored Jun 29, 2015
  
  6bc3c629