louisfd
b2bd0fb313
wip
2024-06-28 14:19:21 -04:00
louisfd
97b53624b9
merged
2024-06-28 11:12:59 -04:00
louisfd
61ca9ff0b6
unroll transpose
2024-06-28 11:08:56 -04:00
louisfd
0dbc3c5af4
in search of the bug
2024-06-28 09:27:05 -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
mepatrick73
1c7780aaac
Feat/dynamic small pool ( #1931 )
2024-06-26 15:42:04 -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
Nathaniel Simard
f9ec2e1006
Handle visibility in cube ( #1929 )
2024-06-26 12:57:47 -04:00
Nathaniel Simard
d772a1cfd5
Fix: launch without generics ( #1932 )
2024-06-26 12:57:32 -04:00
louisfd
d9b4801448
some debugging done
2024-06-26 12:17:27 -04:00
mepatrick73
4c9097030f
Perf/dynamic mm slice adressing ( #1917 )
...
* basic implementation of virtual memory addressing for fast index + merging (there is a bug with slice padding
2024-06-25 18:16:46 -04:00
louisfd
379d15918c
refactoring
2024-06-25 14:45:55 -04:00
Nathaniel Simard
2fbc4628f3
Feat/cube/array assign ops ( #1914 )
2024-06-25 09:55:55 -04:00
Dilshod Tadjibaev
2c51615471
Print model structure like with PyTorch - Part 1 ( #1912 )
2024-06-25 09:23:10 -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
louisfd
673f1468b7
wip
2024-06-20 16:11:16 -04:00
Nathaniel Simard
a5dfb87828
Feat/comptime expr ( #1910 )
...
* Support comptime expressions
* Add test
* Cleanup
* Fix
2024-06-20 16:00:22 -04:00
louisfd
c938374ef0
fix
2024-06-20 14:55:27 -04:00
louisfd
7e3a28a126
tests for computation
2024-06-20 14:52:08 -04:00
louisfd
fc8d3fb087
merge main
2024-06-20 07:31:03 -04:00
louisfd
e7161440d9
wip
2024-06-20 07:05:04 -04:00
Nathaniel Simard
efc13d9a38
Feat/cube/compile error ( #1909 )
2024-06-19 17:21:32 -04:00
Nathaniel Simard
d50bac165e
feat cube support Array ( #1907 )
2024-06-19 17:03:02 -04:00
Arthur Brussee
14d1bbba64
Do not use default burn-compute features unless enabled. ( #1908 )
2024-06-19 10:12:11 -04:00
louisfd
e651e48099
to vectorized not vectorized
2024-06-19 09:50:28 -04:00
louisfd
cebd4d4284
debugging wip
2024-06-18 21:48:53 -04:00
Nathaniel Simard
560d77d154
Doc: Improve module to_device/fork docs ( #1901 )
2024-06-18 16:45:38 -04:00
Nathaniel Simard
e758fd43db
Fix: constant record loading ( #1902 )
2024-06-18 16:45:21 -04:00
Justin Restivo
263add23a0
Tanh nn wrapper ( #1903 )
2024-06-18 16:45:04 -04:00
phenylshima
f8a7c54272
feat: Make RetroForward public ( #1905 )
2024-06-18 16:44:32 -04:00
jachym
96468fc3c9
feat: added reduce min onnx import ( #1894 )
2024-06-18 09:04:24 -04:00
Nathaniel Simard
4f6db974a1
Perf/dynamic mm ( #1906 )
2024-06-18 08:41:07 -04:00