Commit Graph

735 Commits

Author SHA1 Message Date
Mathias Insley 255dfefab2
Feat/tensor unfold (#819) 2023-10-15 17:05:34 -04:00
AuruTus 84e74df3b9
Feat/onnx import erf (#859) 2023-10-13 09:59:44 -04:00
Luni-4 d912f5c5e5
misc: Add clippy lints in PR comments (#659) 2023-10-13 09:57:47 -04:00
Nathaniel Simard e581912c8a
Update book (#858) 2023-10-11 09:11:28 -04:00
Nathaniel Simard 620b86de98
Feat training events (#857) 2023-10-10 13:27:03 -04:00
Dilshod Tadjibaev 097fd956d0
Upgrade dependency versions (#854)
This updates dependencies including tch to 0.14.0, which uses Torch 2.1.
2023-10-09 14:29:44 -04:00
Nathaniel Simard 6c125e161f
Fix CI Rust Update 1.73 (#849) 2023-10-06 12:53:45 -04:00
Nathaniel Simard 904ff1a974
Refactor burn-train (#847) 2023-10-05 13:10:54 -04:00
Eddie Mata Ewy 9afc76303f
removed batch_size field from GRU and LSTM (#845)
* removed batch_size field from GRU and LSTM

* slight style difference

---------

Co-authored-by: Eduardo Mata Ewy <eddiemata@hotmail.es>
2023-10-05 11:15:20 -04:00
Dilshod Tadjibaev e2a17e4295
Add image classification web demo with WebGPU, CPU backends (#840) 2023-10-05 10:29:13 -04:00
Damien Elmes 28e2a99efe
Use a u64 counter for autodiff NodeIDs (#843) 2023-10-04 11:53:34 -04:00
Nathaniel Simard ce120ead3a
Improve metrics (#844) 2023-10-03 18:15:43 -04:00
Nathaniel Simard 809ad72843
Fix flaky init (#842)
* Fix flaky init

* Remove print
2023-10-03 13:57:16 -04:00
Louis Fortier-Dubois 163e48c969
wgpu: Yet another (faster) matmul (#836) 2023-10-02 14:05:53 -04:00
Damien Elmes e363813911
Use thread-local RNG to generate IDs (#839)
We've been exploring dividing our data set up into multiple batches,
and training those batches in parallel. I noticed that performance did
not scale with core count, and after some digging, found that this was
mainly due to the Mutex being used to generate IDs. With the following
change, training across 16 cores went from 21s to 4.2s.

thread_rng was previously discussed on #703, but I don't believe that
applies here, as this is just used for UUID creation.
2023-10-02 14:05:27 -04:00
Daniel Vainsencher 1e609acfe5
Bugfix: epoch and iteration were swapped. (#838)
Co-authored-by: Daniel Vainsencher <danielv@nirinor.ai>
2023-10-01 09:48:51 -04:00
Nathaniel Simard 80731913c7
Speedup CI (#835) 2023-09-29 15:13:16 -04:00
Nathaniel Simard ca787d6446
Feat/async read (#833) 2023-09-28 17:09:58 -04:00
Louis Fortier-Dubois aa90fe8efb
Refactor/burn benchmark (#829) 2023-09-28 09:38:21 -04:00
Alex Errant d06cc2f239
only use gix-tempfile if on sqlite (#830) 2023-09-28 09:22:54 -04:00
Luni-4 6458000b7b
Improve cache key for std checks (#831) 2023-09-26 11:36:43 -04:00
Nathaniel Simard 95e660488e
Refactor/burn compute wgpu (#826) 2023-09-25 10:42:45 -04:00
Arvid Hammarlund 7d706fae98
Added covariance and diagonal ops. (#793) 2023-09-25 08:49:52 -04:00
Dilshod Tadjibaev 07cbd209eb
Improve CI cache - remove burn-tch artifacts (#825) 2023-09-22 11:49:21 -04:00
Dilshod Tadjibaev 3ac9abbacb
Allow custom file location for loading record for ONNX model (#824) 2023-09-22 08:40:36 -04:00
Louis Fortier-Dubois 8c215e8be3
Bugfix/int swap dims (#823) 2023-09-22 08:38:38 -04:00
Dilshod Tadjibaev bdcd5873c5
Support no-std for models generated by burn-import (#822) 2023-09-21 11:43:07 -04:00
Dilshod Tadjibaev 35ca839e18
Add support for different record types in ONNX (#816) 2023-09-21 09:06:57 -04:00
Juliano Decico Negri 293020aae6
#384 Include tests for int.rs and float.rs (#794) 2023-09-21 09:00:09 -04:00
Dilshod Tadjibaev 393d86e99d
Fix dataset count slowness (#818)
Fixes #811
2023-09-21 08:55:13 -04:00
Nathaniel Simard aacf191161
Fix training checkpoints (#815) 2023-09-21 08:52:04 -04:00
Nathaniel Simard ac4adb54ea
Burn compute (#809) 2023-09-18 19:56:53 -04:00
Damien Elmes d7e9e75099
Fix train-minimal feature and ensure it gets tested (#802) 2023-09-16 09:52:14 -04:00
Peter Szilvasi 6c32e23f15
Fix the device initialization in backend.md (#808) 2023-09-16 09:44:34 -04:00
Luni-4 5691faa3c2
ci: Improve cache key to better discriminate among caches (#804) 2023-09-15 09:42:49 -04:00
Arvid Hammarlund 8f65acf65a
Made docs show features not included in std. (#789) 2023-09-15 09:14:44 -04:00
Nathaniel Simard 57d6a566be
Feat/dashboard tui (#790) 2023-09-13 10:45:14 -04:00
Nathaniel Simard 4f72578260
Fix compilation flags (#800) 2023-09-13 10:38:16 -04:00
Damien Elmes 6187d88226
Check examples when all tests are run (#801)
Will break CI until #800 is merged.
2023-09-13 09:56:56 -04:00
Nathaniel Simard 251ec00070
Fix clamp API (#799) 2023-09-12 14:20:11 -04:00
Damien Elmes dbc5f0a340
Fix missing train re-export when using train-minimal (#785)
Accidental regression in #773
2023-09-08 09:04:09 -04:00
Justin Moore d2278002e4
Avoid building `xtask` when invoking `cargo build` (#784) 2023-09-08 09:01:39 -04:00
Louis Fortier-Dubois d249d24c07
simplify getting started (#782) 2023-09-07 14:57:32 -04:00
Dilshod Tadjibaev 27952b1f47
Fix broken link and other minor fixes (#780) 2023-09-07 08:44:57 -04:00
Luni-4 c0eb339a8e
Remove binaries from .gitignore (#775) 2023-09-07 08:43:52 -04:00
Nathaniel Simard af0be5cfeb
Chore: bump version (#777) 2023-09-06 12:15:13 -04:00
nathaniel 76894eff04 Remove candle from release 2023-09-06 10:25:37 -04:00
Nathaniel Simard 8be5889367
Prepare next release (#774)
* Prepare for next release

* Fix
2023-09-06 09:29:05 -04:00
Arvid Hammarlund f330f9f73d
Added Dim-iter (#771)
* "added dim-iter".

* added tests to iter_dim

* Made the rnn implemenations use squeeze.
2023-09-06 09:22:30 -04:00
Nathaniel Simard 8b3d10c4d3
Cleanup of the book + feature flags (#773)
---------

Co-authored-by: louisfd <louisfd94@gmail.com>
2023-09-06 09:16:36 -04:00