Commits · c37e6e1f127504e85b372f4d52bfc149709000dc · OpenDAS / torchani

30 Aug, 2018 1 commit
- Inference benchmark (#81) · c37e6e1f
  Gao, Xiang authored Aug 30, 2018
  
  c37e6e1f
28 Aug, 2018 1 commit
- temporarily disable l2 support for now, planning to reenable it after adaw get... · 0ed9b3f3
  Gao, Xiang authored Aug 28, 2018
```
temporarily disable l2 support for now, planning to reenable it after adaw get merged into pytorch (#80)
```
  0ed9b3f3
27 Aug, 2018 2 commits
- test network architecture and trainer setups (#79) · bfa36346
  Gao, Xiang authored Aug 27, 2018
  
  bfa36346
- L2 only on weight · d673f672
  Xiang Gao authored Aug 27, 2018
  
  d673f672
26 Aug, 2018 1 commit
- NeuroChem trainer loader (#78) · 81c163e7
  Gao, Xiang authored Aug 26, 2018
  
  81c163e7
25 Aug, 2018 1 commit
- Docs improvements (#77) · 9cae6d3f
  Gao, Xiang authored Aug 24, 2018
  
  9cae6d3f
23 Aug, 2018 1 commit
- Improve training related API (#76) · 3cced1e6
  Gao, Xiang authored Aug 23, 2018
  
  3cced1e6
22 Aug, 2018 3 commits
- Remove energy shifter's dependency on atomic symbols (#75) · d7ef8182
  Gao, Xiang authored Aug 22, 2018
  
  d7ef8182
- simplify handling of species to avoid unnecessary arguments (#74) · eb090700
  Gao, Xiang authored Aug 22, 2018
  
  eb090700
- Some API improvements (#73) · ce21d224
  Gao, Xiang authored Aug 21, 2018
  
  ce21d224
21 Aug, 2018 3 commits
- Refactor codes to put neurochem related codes together (#72) · f146feca
  Gao, Xiang authored Aug 21, 2018
  
  f146feca
- update readme (#70) · 85a6dd1e
  Gao, Xiang authored Aug 21, 2018
  
  85a6dd1e
- remove dependency on sphinx on setup.py (#69) · 61637872
  Gao, Xiang authored Aug 21, 2018
  
  61637872
19 Aug, 2018 1 commit
- fix energy force example (#67) · f4c703bb
  Gao, Xiang authored Aug 19, 2018
  
  f4c703bb
17 Aug, 2018 1 commit
- Split batch to avoid performance penalty on padding (#66) · 99c2c3dc
  Gao, Xiang authored Aug 17, 2018
  
  99c2c3dc
13 Aug, 2018 3 commits
- fix padding strip when batching (#65) · 22975fa7
  Gao, Xiang authored Aug 13, 2018
  
  22975fa7
- preprocess data on CPU to save cuda memory (#64) · 7c9402d0
  Gao, Xiang authored Aug 13, 2018
  
  7c9402d0
- Modify training related APIs to batch the whole data by padding (#63) · 50fb6cdb
  Gao, Xiang authored Aug 13, 2018
  
  50fb6cdb
11 Aug, 2018 1 commit
- Support batching non isomers in models and aev computer (#62) · b7cab4f1
  Gao, Xiang authored Aug 11, 2018
  
  b7cab4f1
10 Aug, 2018 2 commits
- add TransformedLoss to allow adding non-linearity to computed loss (#61) · b63f6e40
  Gao, Xiang authored Aug 10, 2018
  
  b63f6e40
- use more specific ops for better performance (#60) · b16a3234
  Gao, Xiang authored Aug 09, 2018
  
  b16a3234
08 Aug, 2018 3 commits
- add support for early stopping (#59) · 7606fc3b
  Gao, Xiang authored Aug 08, 2018
  
  7606fc3b
- Rename to better describe what it is (#58) · f93a9ba4
  Gao, Xiang authored Aug 08, 2018
  
  f93a9ba4
- fix sort_by_species (#57) · 5bb66915
  Gao, Xiang authored Aug 07, 2018
  
  5bb66915
07 Aug, 2018 2 commits
- a new example to test neurochem on specified dataset (#56) · e5439f3d
  Gao, Xiang authored Aug 07, 2018
  
  e5439f3d
- log training rmse and best validation rmse · 7f660099
  Gao, Xiang authored Aug 07, 2018
  
  7f660099
03 Aug, 2018 5 commits
- checkpoint if validation gets better result (#55) · 3e296069
  Gao, Xiang authored Aug 03, 2018
  
  3e296069
- remove output_length (#54) · f50cc0b4
  Gao, Xiang authored Aug 03, 2018
  
  f50cc0b4
- improve energy shifter (#52) · 73e447f0
  Gao, Xiang authored Aug 03, 2018
  
  73e447f0
- implement per atom loss (#51) · 59b31d84
  Gao, Xiang authored Aug 03, 2018
  
  59b31d84
- merge Container and BatchModel to provide richer information in output (#50) · 84439caf
  Gao, Xiang authored Aug 03, 2018
  
  84439caf
02 Aug, 2018 3 commits
- add argparser for nnp_training.py (#46) · 1d8bba37
  Gao, Xiang authored Aug 02, 2018
  
  1d8bba37
- add argparse to training-benchmark.py (#45) · 861580e3
  Gao, Xiang authored Aug 02, 2018
  
  861580e3
- remove explicit device and dtype (#44) · d3ae0788
  Gao, Xiang authored Aug 02, 2018
  
  d3ae0788
01 Aug, 2018 2 commits
- use 0 size n dim tensor feature to simplify code (#43) · 8c493a6e
  Gao, Xiang authored Aug 01, 2018
  
  8c493a6e
- use pytorch's CELU to simplify code (#42) · 0e992fe5
  Gao, Xiang authored Aug 01, 2018
  
  0e992fe5
31 Jul, 2018 4 commits
- fix build (#41) · 97e3df07
  Gao, Xiang authored Jul 31, 2018
  
  97e3df07
- completely decouple aev computer and neural networks (#40) · e2e740d0
  Gao, Xiang authored Jul 31, 2018
  
  e2e740d0
- remove force from ANIModel (#39) · ee31b752
  Gao, Xiang authored Jul 31, 2018
  
  ee31b752
- make aev computer accept species as tensor (#38) · 2cb357bd
  Gao, Xiang authored Jul 31, 2018
  
  2cb357bd