dnrops
/
burn
mirror of https://github.com/tracel-ai/burn.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
488 Commits 49 Branches 17 Tags 90 MiB
Commit Graph

103 Commits

Author SHA1 Message Date
Wouter Doppenberg 4c663b4cb7
Added correct expected weights, fixed adam state init (#621) 2023-08-09 17:55:39 -04:00
Caio Piccirillo cb283a9e5b
Max pool1d (#602) 2023-08-09 16:13:48 -04:00
Wouter Doppenberg 1f01fcb640
AdamW implementation (#566) 2023-08-09 16:09:17 -04:00
Dilshod Tadjibaev 716b7569de
Add Dropout ONNX OP and other refactoring (#606) 2023-08-09 09:53:20 -04:00
Caio Piccirillo 1d3bbaab13
Typos (#608) 2023-08-08 17:57:51 -04:00
Gadersd ed255c5561
Use buffered io for massive performance gains when loading and saving… (#593) 2023-08-06 12:56:27 -04:00
Nathaniel Simard ce8a175aa4
Feat/conv transpose1d backward (#586) 2023-08-06 10:50:10 -04:00
Dilshod Tadjibaev 1554a3c898
Full support for ONNX scalar operators and Constants (#578) 2023-08-04 16:51:51 -04:00
Nathaniel Simard ca9a8808d9
Feat/adaptive avg pool1d (#585) 2023-08-04 13:55:18 -04:00
Nathaniel Simard 8436d4ff66
Feat/tensor/adaptive avg pool2d (#572) 2023-08-04 10:23:59 -04:00
Nathaniel Simard 597eab524d
Feat/conv transpose2d (#574) 2023-08-03 15:42:18 -04:00
Nathaniel Simard 0a5a2d729a
chore: bump version for next release (#533) 2023-07-26 09:46:28 -04:00
Luni-4 914820e6e4
burn-core: Use a specific structure for 1d padding (#529) 2023-07-24 19:31:37 -04:00
polina guseva 64090a582b
Add static full method for tensor intialization with custom values (#486) 
---------

Co-authored-by: Dilshod Tadjibaev <939125+antimora@users.noreply.github.com>
 2023-07-24 13:03:27 -04:00
Luni-4 e066d95d2e
Implement padding for conv2d (#523) 2023-07-24 12:54:45 -04:00
Luni-4 f70ff4dc54
Add MaxPool2d to burn-import (#507) 2023-07-23 19:28:36 -04:00
Louis Fortier-Dubois 57a5476c89
bugfix for macos test (#503) 2023-07-18 16:15:00 -04:00
Dilshod Tadjibaev f433292a3f
Fix intermittent test failure (#497) 2023-07-13 19:40:54 -04:00
Dilshod Tadjibaev 53c088209d
Fix new clippy warnings that cause the CI to fail (#494) 2023-07-13 13:39:39 -04:00
Nathaniel Simard 4bc740a864
Fix: can't use init or load with a record with correct gradient requi… (#488) 2023-07-11 15:01:14 -04:00
Dilshod Tadjibaev 20213ade73
Add sinusoidal positional embedding module (#481) 2023-07-11 10:55:29 -04:00
Dilshod Tadjibaev e62ee1269b
Fix burn-tch's random implementation for standard dist (#469) 2023-07-06 08:50:50 -04:00
Nathaniel Simard 65bf6c1cbb
Refactor index => slice (#466) 2023-07-05 16:30:11 -04:00
Dilshod Tadjibaev a2b2d99aa3
Add Tensor as a module constant (#459) 2023-07-03 13:00:18 -04:00
Nathaniel Simard 5341c0260b
Perf/wgpu/benches (#448) 2023-07-02 12:57:15 -04:00
Dilshod Tadjibaev aced634827
Remove hardcoded "1" stride from Conv1d (#451) 2023-07-01 10:31:16 -04:00
Nathaniel Simard 13836ac599
Fix: wgpu powf (#443) 2023-06-28 12:50:16 -04:00
Dilshod Tadjibaev 61291cc30c
Revert temp half-rs fix for no_std (#430) 
Fixes #268
 2023-06-26 15:59:06 -04:00
Dilshod Tadjibaev eda241f8cf
Add missing docs and enable missing_docs warn lint (#420) 2023-06-21 14:12:13 -04:00
Yu Sun 9c5afa3469
fix(loss): Declare all contents of reduction as public (#423) 2023-06-21 13:00:57 -04:00
Dilshod Tadjibaev fce45f51be
Doc fixes (#418) 2023-06-21 12:32:50 -04:00
Dilshod Tadjibaev 4683acf726
Minor clean up of doc formatting and remove outdated TODO (#415) 2023-06-20 10:05:28 -04:00
Louis Fortier-Dubois e5cc50a837
refactor/initializer (#398) 
* add tests to linear forward

* swap linear weight dims

* rename uniformdefault

* wip bias in init

* refactor initializer

* formatting

* fix get(0) to first

* wip refactor

* builder pattern + kaiming init

* kaiming initializer and major initializer refactor

* fix fan out

* easier fixes

* remove initializer options

* clippy fix

* revert to input*output linear

* gru uses xavier init

* should be normal actually

---------

Co-authored-by: louisfd <louisfortier-dubois@Louiss-MacBook-Pro.local>
 2023-06-19 16:55:11 -04:00
Dilshod Tadjibaev 834c7ecc1f
Clean up cargo descriptions and formatting (#403) 2023-06-15 09:20:53 -04:00
Louis Fortier-Dubois fd88398ce4
feat/Xavier glorot (#394) 2023-06-11 09:54:44 -04:00
Mathias Insley f8cd38c071
Feat/gru (#393) 2023-06-09 18:56:40 -04:00
Mathias Insley 8a88a868ee
Feat/lstm (#370) 2023-06-06 14:33:22 -04:00
Yu Sun 498d163a7b
Feat: mse loss (#378) 2023-06-03 10:31:12 -04:00
Nathaniel Simard 974fdfaba1
Feat/wgpu backend setup (#376) 2023-06-02 11:52:47 -04:00
Nathaniel Simard 2a4ba5a6ab
Feat/gather scatter (#367) 2023-05-27 11:40:04 -04:00
Nathaniel Simard 3ef2a18d87
Fix flaky tests + add feature flag (#362) 2023-05-24 08:33:31 -04:00
Nathaniel Simard ba0ac112e1
Refactor: Burn Import use BurnGraph (#355) 2023-05-21 13:09:27 -04:00
Nathaniel Simard 18cb19cd03
Fix deserializer array of constant (#354) 2023-05-19 12:01:59 -04:00
Nathaniel Simard 2df268b674
feat: improve module recorder deserialization (#351) 2023-05-15 08:29:57 -04:00
Nathaniel Simard 976102fec0
Feat/avg pool1d (#349) 2023-05-15 08:29:45 -04:00
Dilshod Tadjibaev 05763e1878
Bump version to the next minor to indicate dev (#344) 2023-05-10 18:02:08 -04:00
Nathaniel Simard 29eecd6383
Prepare next release (#335) 2023-05-06 10:32:23 -04:00
Nathaniel Simard a50262b730
refactor: grad clipping (#326) 2023-05-05 12:56:35 -04:00
Mathias Insley d892f6186c
Gradient Clipping (#305) 2023-05-04 15:01:30 -04:00
Nathaniel Simard 6c9a5c8e58
Refactor/record (#323) 2023-05-04 14:59:16 -04:00