dnrops
/
burn
mirror of https://github.com/tracel-ai/burn.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
1,293 Commits 39 Branches 19 Tags 101 MiB
Commit Graph

126 Commits

Author SHA1 Message Date
louisfd f7f5722632 cleanup 2024-07-04 14:21:14 -04:00
louisfd e87c553c9a benchmark 2024-07-04 14:16:00 -04:00
louisfd 43877da1f2 offset check bounds work 2024-07-04 13:42:26 -04:00
louisfd cbb4de7156 refactor load shared memory file 2024-07-04 10:52:54 -04:00
louisfd 74b06fbf55 transposed fails on purpose 2024-07-04 10:25:04 -04:00
louisfd 9d61e680f3 jit layout info refactor 2024-07-04 10:11:49 -04:00
louisfd 2292b38778 comptime else if 2024-07-03 15:54:25 -04:00
louisfd 016456b0d9 clippy 2024-07-03 14:08:45 -04:00
louisfd c4cf3f33cf typo 2024-07-03 13:58:23 -04:00
louisfd 60b9b9fc84 clean tests 2024-07-03 13:56:06 -04:00
louisfd f875793f25 ignore clippy 2024-07-03 13:52:58 -04:00
louisfd 657ad60b14 ignore clippy 2024-07-03 13:52:25 -04:00
louisfd 9773cad04b ignore clippy 2024-07-03 13:52:00 -04:00
louisfd 90bfacad79 fmt 2024-07-03 13:47:54 -04:00
louisfd 623e7a22a4 refactoring 2024-07-03 13:46:44 -04:00
louisfd ed86d91e7a several unrolling options 2024-07-03 12:58:48 -04:00
louisfd d6cee6c402 merge main 2024-07-03 12:40:19 -04:00
Nathaniel Simard 679cfd6dfb
Refactor cube launch + support inplace operation (#1961) 2024-07-03 11:58:35 -04:00
Guillaume Charifi d696d74e3d
Implement 3D and transposed 3D convolutions. (#1945) 
* Implement 3D and transposed 3D convolutions.

* Merge changes from onnx-ir #1921 pr

---------

Co-authored-by: Dilshod Tadjibaev <939125+antimora@users.noreply.github.com>
 2024-07-02 17:54:35 -05:00
louisfd 173374ef95 can make empty vector 2024-07-02 13:02:18 -04:00
louisfd 34b87fc5b4 check sm bounds 2024-07-02 11:06:17 -04:00
louisfd 71cc377793 upstream info access 2024-07-02 10:20:06 -04:00
louisfd 0c0a357746 merge main 2024-07-02 08:37:23 -04:00
Nathaniel Simard 82a883a57d
Feat/cube/fma (#1947) 2024-07-02 08:32:39 -04:00
Nathaniel Simard cb6b5e7183
Feat/cube/cooperative matrix-multiply and accumulate. (#1943) 2024-07-02 08:31:00 -04:00
Arthur Brussee 849c8f453b
Consistent sync/async handling, allow more functions to be async for wasm. (#1936) 2024-07-02 08:25:28 -04:00
louisfd 61ca9ff0b6 unroll transpose 2024-06-28 11:08:56 -04:00
louisfd 081fd782af fix unroll bug 2024-06-27 16:59:50 -04:00
louisfd 71ea9fb415 benchmarked, it's slower :( 2024-06-27 16:55:04 -04:00
louisfd 3b473597df &mut in matmul 2024-06-27 16:42:21 -04:00
Nathaniel Simard 1ae1c03b2d
Refactor/cube/mutability (#1934) 2024-06-27 16:03:23 -04:00
louisfd 32c2c1b292 fix vectorization bug 2024-06-27 16:01:23 -04:00
louisfd a3c081ac7d wip merge main 2024-06-27 15:39:27 -04:00
louisfd e20a5f0a81 matmul cube tests refactor 2024-06-27 15:09:05 -04:00
louisfd 4e82ad410c cleaner load shared tests 2024-06-27 14:34:18 -04:00
louisfd 55d4e0d886 minor cleanup 2024-06-27 12:56:05 -04:00
louisfd 1746d41358 refactor write for comptime n checks 2024-06-27 11:32:01 -04:00
louisfd 6b58882b40 refactor for horizontal check comptime 2024-06-27 10:00:22 -04:00
louisfd 5d9646bb1b works everytime 2024-06-27 07:57:58 -04:00
Guillaume Lagrange cdd1fa1672
Refactor tensor data (#1916) 
* Move distribution to module

* Add new TensorData with serialization support

* Implement display and from for TensorData

* Add missing Cargo.lock

* Add missing bytemuck feature

* Add zeros, ones, full and random TensorData methods

* Refactor Data -> TensorData usage

* Fix tests

Since TensorData is not generic over the element type anymore no type inference can be done by the compiler. We must explicitly cast the expected results to the expected backend type.

* Remove commented line

* Fix import

* Add record-backward-compat

* Remove dim const generic from TensorData

* Support NestedValue de/serialization with TensorData

* Fix burn-jit tests

* Remove eprinln

* Refactor onnx import to use TensorData

* Fix tch from_data

* Fix nested value serialization for u8

* Fix missing import

* Fix reduce min onnx test

* Fix deprecated attribute

* Remove shape getter

* Remove strict assert in tests

* Add tensor data as_bytes

* Add tensor check for rank mismatch

* Fix typo (dimensions plural)

* Fix error message

* Update book examples with from_data and fix Display impl for TensorData

* Add deprecation note
 2024-06-26 20:22:19 -04:00
louisfd 4cb4f22599 more wip 2024-06-26 18:33:20 -04:00
louisfd b93e984555 wip fixing last bug 2024-06-26 17:32:53 -04:00
louisfd 122b909e75 variable vectorization for output 2024-06-26 15:16:30 -04:00
louisfd 15483bed4f works 2024-06-26 14:16:24 -04:00
louisfd d9b4801448 some debugging done 2024-06-26 12:17:27 -04:00
louisfd 379d15918c refactoring 2024-06-25 14:45:55 -04:00
louisfd 17a01d0131 refactoring 2024-06-22 13:25:59 -04:00
louisfd 55c117874a tested write output 2024-06-21 14:48:31 -04:00
louisfd 1404327e70 read tile tests 2024-06-21 08:40:00 -04:00
louisfd 40a874cdcd merge main 2024-06-20 16:28:10 -04:00