Louis Fortier-Dubois
acfd1e0fe9
Doc/book/getting started ( #744 )
2023-08-31 14:08:13 -04:00
Dilshod Tadjibaev
510d2d4cb3
Apply markdown formatting using prettier ( #743 )
2023-08-31 11:44:29 -04:00
Nathaniel Simard
c95b34c511
Book: backend extension + custom wgpu kernel ( #728 )
2023-08-31 09:55:43 -04:00
Caio Piccirillo
0dd4ef6b75
Add configuration doc for vscode environment setup ( #737 )
2023-08-31 08:23:28 -04:00
Louis Fortier-Dubois
760c9e1d8e
Feat/candle/module ops ( #725 )
2023-08-30 18:53:03 -04:00
Nathaniel Simard
aafceeffa0
Refactor: Module & Record derive ( #735 )
2023-08-30 12:42:49 -04:00
Dilshod Tadjibaev
40a6573f28
Book: Expanded the planned sections and added built-in module section ( #730 )
2023-08-30 12:39:35 -04:00
Louis Fortier-Dubois
f253f19b4e
add tanh ( #733 )
2023-08-30 10:00:50 -04:00
Louis Fortier-Dubois
7c34e21424
Perf/tensor ops/more tests ( #718 )
2023-08-30 09:08:18 -04:00
Dilshod Tadjibaev
bb71f17409
Update burn-import README ( #727 )
2023-08-30 08:20:53 -04:00
Nathaniel Simard
4ea96627a8
Disable fake dataset by default ( #720 )
2023-08-29 12:04:31 -04:00
Damien Elmes
0fd82ee5d3
Allow opting out of SQLite's bundling, or all SQLite usage ( #711 )
2023-08-29 10:50:01 -04:00
Damien Elmes
ff1c0d8f1a
Fix: ensure final CLI update happens ( #716 )
...
The merge of #708 unearthed a bug in the CLI code: if at completion time
the update is within the throttling period, you can end up with a final
output that appears as if the process didn't fully complete.
More info: https://github.com/open-spaced-repetition/fsrs-optimizer-burn/pull/36#issuecomment-1696736807
2023-08-29 07:59:12 -04:00
Dilshod Tadjibaev
13342b45d1
Account for additional cases in Linear node conversion ( #709 )
2023-08-28 16:24:59 -04:00
Damien Elmes
a4a9844da3
Feat: Some tweaks to make it more practical to integrate in a GUI app ( #706 )
...
* feat: Add support for using a custom renderer
When integrating in an app, the CLI display is undesirable. This will
allow us to collect the progress of iterations, so they can be displayed
in a GUI.
Because CLIDashboardRenderer() writes to the console when ::new() is
called, the code has had to be refactored to defer creation until .build()
is called. This meant that instead of delegating the metric assignments
to the already-created dashboard, we instead need to store them and add
them later.
* feat: Allow opt-out of experiment.log
2023-08-28 16:23:31 -04:00
Nathaniel Simard
968cd6e390
Feat: add dataset sampler with and without replacement ( #704 )
2023-08-28 13:08:05 -04:00
Louis Fortier-Dubois
c89f9969ed
Perf/tensor ops/tests ( #710 )
2023-08-28 12:53:17 -04:00
Nathaniel Simard
f024dc9ccb
fix: dataloader reshuffle dataset each iter ( #703 )
2023-08-28 12:48:33 -04:00
Justin Moore
3aa7a3318e
Optimize `argmin`/`argmax` ( #707 )
2023-08-28 12:15:56 -04:00
Damien Elmes
9f1a68bc33
Fix: processed count was one greater than total items at completion ( #708 )
2023-08-28 11:56:42 -04:00
Mathias Insley
d2aa4c0c9d
Perf/Empty Context Cache ( #676 )
...
* Add a pipeline_counter and methods for process of retaining best kernel
* Put a tune flag on the Context
* Put counts into cache instead of using pipeline_counter
* Formatting
* Add optimize_cache flag and rework ComputePipeline clearing process
* Update tune() so that it starts Context tuning and flags the Context as ready for clearing
* Consistent single quotes
* Use AtomicBool for is_tuning, prevent caching during tuning
* Collect TemplateIds during tuning and clean them out after tuning
* Fix comment
* Move cache cleanup to stop_tuning function
2023-08-28 10:04:05 -04:00
Nathaniel Simard
084c8bb4e0
Fix: autodiff backward broadcast ( #702 )
2023-08-28 08:20:38 -04:00
Asuka Minato
0f7864f11b
Path type ( #699 )
2023-08-27 09:17:48 -04:00
MOZGIII
7f558bdc46
Expose element traits ( #700 )
2023-08-27 09:02:39 -04:00
Dilshod Tadjibaev
03ee987cda
ONNX Clip Operator ( #696 )
2023-08-26 17:58:51 -04:00
Nathaniel Simard
a25f8b224a
Fix: grad replace was adding to previous tensor ( #695 )
2023-08-25 14:58:18 -04:00
AuruTus
3fdfeeae2b
Feat/onnx import tanh ( #683 )
2023-08-25 14:57:07 -04:00
Louis Fortier-Dubois
fb2a71bb81
remove to device ( #694 )
2023-08-25 09:55:18 -04:00
Will Brickner
bab036b263
Added custom initializer option for MHA and other transformer-related modules ( #693 )
2023-08-25 09:55:05 -04:00
Louis Fortier-Dubois
88cb6b07fc
Feat/candle/more operations ( #682 )
2023-08-25 08:46:30 -04:00
Nathaniel Simard
481ff14fe1
feat: can add custom training and validation metric loggers ( #690 )
2023-08-25 07:16:45 -04:00
Nathaniel Simard
efee0ac296
Feat/train/custom optimize method ( #689 )
...
* Add the possibility to add a custom optimize function for models
* Fix clippy
2023-08-25 07:14:36 -04:00
Nathaniel Simard
183620fb20
feat: replace grad tensor ( #688 )
2023-08-24 14:07:04 -04:00
Jerome Robert
edb3e9fc4b
Do not use default device when running kernel::matmul::tune ( #684 )
...
Use the device of the involved Tensor instead of Device::default
2023-08-24 14:01:27 -04:00
Louis Fortier-Dubois
bc27a87e9d
new clippy stuff ( #687 )
2023-08-24 13:20:58 -04:00
Mahmoud
b263faebc3
fix(book): use the correct burn versions ( #681 )
2023-08-23 14:07:25 -04:00
Nathaniel Simard
d18d1b0bb9
Can configure wgpu max tasks ( #603 )
2023-08-23 12:20:27 -04:00
Nathaniel Simard
00d3d208b8
Add book + fix some code ( #671 )
2023-08-23 11:52:55 -04:00
Mahmoud
b60d931771
chore: update README.md ( #678 )
2023-08-23 10:32:51 -04:00
dependabot[bot]
8f4faabd5f
Update gix-tempfile requirement from 7.0.0 to 8.0.0 ( #679 )
...
Updates the requirements on [gix-tempfile](https://github.com/Byron/gitoxide ) to permit the latest version.
- [Release notes](https://github.com/Byron/gitoxide/releases )
- [Changelog](https://github.com/Byron/gitoxide/blob/main/CHANGELOG.md )
- [Commits](https://github.com/Byron/gitoxide/commits )
---
updated-dependencies:
- dependency-name: gix-tempfile
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2023-08-23 10:32:30 -04:00
Will Brickner
d1a708e317
Made TchTensor fields public ( #677 )
2023-08-22 21:37:58 -04:00
Caio Piccirillo
2fefc82099
Dilation maxpool ( #668 )
2023-08-21 14:14:25 -04:00
Elazrod56
dd5ea5251c
Training metrics ( #647 )
2023-08-21 14:13:36 -04:00
Louis Fortier-Dubois
b07af74788
support broadcast matmul ( #669 )
2023-08-21 11:43:21 -04:00
Dilshod Tadjibaev
b06fe660e8
Onnx tests and bug fixes - BatchNorm, Identity, Relu, Sigmoid and Transpose ( #661 )
2023-08-21 11:39:59 -04:00
Baye Dieng
a557caa4f7
remove channel field from maxpool1d config struct ( #666 )
2023-08-20 19:24:24 -04:00
Louis Fortier-Dubois
6a5ea0ef7c
Feat/candle/basic operations ( #664 )
2023-08-20 18:55:14 -04:00
nathaniel
6b5ba77084
Fix build
2023-08-17 11:13:59 -04:00
Nathaniel Simard
bda03c6a76
Feat/avg pool/include pad config ( #653 )
2023-08-17 08:50:31 -04:00
Louis Fortier-Dubois
c1eddf04fc
Feat/candle/initialize ( #650 )
2023-08-17 08:50:08 -04:00