Commits · 06c26be3c471ab5afec9915db50b9b2058c6b020 · Jon Lachmann / torch · GitLab

Feb 04, 2021
- fix leaks with doubles · 06c26be3
  Daniel Falbel authored Feb 04, 2021
  
  06c26be3
Feb 03, 2021
- autocast for memory format · 28cc4734
  Daniel Falbel authored Feb 03, 2021
  
  28cc4734
- use namespace env · f2342353
  Daniel Falbel authored Feb 03, 2021
  
  f2342353
- generator autocast · 313e0012
  Daniel Falbel authored Feb 03, 2021
  
  313e0012
- Dimname and DimnameList autocast · 80c8fde7
  Daniel Falbel authored Feb 03, 2021
  
  80c8fde7
- dtype autocasting · a0427b21
  Daniel Falbel authored Feb 02, 2021
  
  a0427b21
Feb 02, 2021
- Merge branch 'master' of https://github.com/mlverse/torch into refactor-argument-wrapping · 662e6198
  Daniel Falbel authored Feb 02, 2021
  
  662e6198
- autocast for torch_device's · ef7a8f20
  Daniel Falbel authored Feb 02, 2021
  
  ef7a8f20
- update CUDA support · 6ae289dc
  Daniel Falbel authored Feb 02, 2021
  
  6ae289dc
Feb 01, 2021
- Update NEWS.md · f228efdc
  Daniel Falbel authored Feb 01, 2021
  
  f228efdc
- Merge pull request #458 from mlverse/save-compat · ffa5b8ec
  Daniel Falbel authored Feb 01, 2021
  ffa5b8ec
- cant reproduce error on windows... something related to the download. skipping. · 43f4e1b8
  Daniel Falbel authored Feb 01, 2021
  
  43f4e1b8
- download in binary mode · 5e0abd65
  Daniel Falbel authored Feb 01, 2021
  
  5e0abd65
- revert changes from other PR · 96971d5c
  Daniel Falbel authored Feb 01, 2021
  
  96971d5c
- allow loading models saved with older torch versions and add a `version` field to the saved object. · 0b6bba90
  Daniel Falbel authored Feb 01, 2021
  
  0b6bba90
- autocast for tensor options · ac526c83
  Daniel Falbel authored Jan 31, 2021
  
  ac526c83
Jan 31, 2021
- autocast tensor options from Rcpp · 43bd8841
  Daniel Falbel authored Jan 31, 2021
  
  43bd8841
- update description · ec6a045c
  Daniel Falbel authored Jan 31, 2021
  
  ec6a045c
- tensor list auto casting · f908b2be
  Daniel Falbel authored Jan 31, 2021
  
  f908b2be
Jan 30, 2021
- remove install copy as we might be running out of storage space · fcb07d48
  Daniel Falbel authored Jan 30, 2021
  
  fcb07d48
- handle NULL -> XPtrTorchScalar and add regression test · 137d82f9
  Daniel Falbel authored Jan 30, 2021
  
  137d82f9
- support converts from SEXP's to XPtrTorchScalar · de56c529
  Daniel Falbel authored Jan 30, 2021
  
  de56c529
- Merge branch 'master' of https://github.com/mlverse/torch into refactor-argument-wrapping · bd967e23
  Daniel Falbel authored Jan 30, 2021
  
  bd967e23
- Update NEWS.md · e8dd12a5
  Daniel Falbel authored Jan 30, 2021
  
  e8dd12a5
- Merge pull request #456 from mlverse/cuda-allocator · 141aab6a
  Daniel Falbel authored Jan 30, 2021
  141aab6a
- unify main thread ID · bc2fb2bd
  Daniel Falbel authored Jan 29, 2021
  
  bc2fb2bd
Jan 29, 2021
- try back gc from base · 8206a617
  Daniel Falbel authored Jan 29, 2021
  
  8206a617
- dont add cuda allocator by default · 0d229ca5
  Daniel Falbel authored Jan 29, 2021
  
  0d229ca5
- better way to conditionally compile the cuda wrapper · b7ab5cde
  Daniel Falbel authored Jan 29, 2021
  
  b7ab5cde
- better way to selectively include file · 4276dd0a
  Daniel Falbel authored Jan 29, 2021
  
  4276dd0a
- include files conditionally · 695dca43
  Daniel Falbel authored Jan 29, 2021
  
  695dca43
- add unit texst · ffdc8701
  Daniel Falbel authored Jan 29, 2021
  
  ffdc8701
- call r garbage collector when more memory is requested and we are in the main thread. · 547a22ad
  Daniel Falbel authored Jan 29, 2021
  
  547a22ad
Jan 28, 2021
- bootstrap the logic to call the custom callback · 11a0774a
  Daniel Falbel authored Jan 28, 2021
  
  11a0774a
- get started · 7fbbe2eb
  Daniel Falbel authored Jan 28, 2021
  
  7fbbe2eb
Jan 26, 2021
- auto convert atomic vectors into tensors from R. we use an R function it would... · 7fea4a96
  Daniel Falbel authored Jan 26, 2021
  7fea4a96
- also automatically convert NULL into an undefined tensor. · 8ece29d4
  Daniel Falbel authored Jan 26, 2021
  
  8ece29d4
- take XPtrTorchTensors directly and make use of the newly added SEXP constructor. · 649cb12b
  Daniel Falbel authored Jan 26, 2021
  
  649cb12b
- allows unwrapping arguments passed from R to XPtrTorchTensor directly · 6d1cce4c
  Daniel Falbel authored Jan 26, 2021
  
  6d1cce4c
- Update NEWS.md · 480a0113
  Daniel Falbel authored Jan 26, 2021
  
  480a0113