Commit Graph

389 Commits

Author SHA1 Message Date
Daniel Strobusch
7365d142a3
random.uniform must respect dtype, even if lower precision than "low" (#280)
Fix an edge case where random uniform returns a float32 array, even if a lower precision dtype is wanted due to adding the float32 "low" array.
2023-12-24 07:04:43 -08:00
Awni Hannun
8b227fa9af
fix no metal build (#276) 2023-12-23 19:18:10 -08:00
Ronan Collobert
cd3616a463
Revisit autorelease memory pools (#260)
* make general autorelease pool part of metal device

* make things simpler

* no metal backend support

* new_memory_pool -> new_scoped_memory_pool
2023-12-22 11:01:26 -08:00
Awni Hannun
2118c3dbfa
fix (#255) 2023-12-21 18:18:41 -08:00
Awni Hannun
a002797d52
A temporary fix (#254) 2023-12-21 17:59:15 -08:00
Daniel Strobusch
794feb83df
support arange for bfloat16 (#245) 2023-12-21 14:33:43 -08:00
Angelos Katharopoulos
b3916cbf2b
Improve names of quantization arguments (#235)
* Change the default quantization group_size to 64
* Rename groups to group_size and width to bits
2023-12-20 16:53:53 -08:00
Angelos Katharopoulos
57fe918cf8
Adds C++ and nn quantization utilities (#230)
* Add C++ de-/quantize ops
* Add quantize functions to the docs and tests
* Add a QuantizedLinear module
2023-12-20 14:17:38 -08:00
Angelos Katharopoulos
2807c6aff0
Implements divide for integer types and adds floor_divide op (#228)
* Add floor_divide
* Add floor_divide to the tests
* Add floor_divide to the docs
2023-12-19 20:12:19 -08:00
davidkoski
de892cb66c
fix for non-macos build issue on cblas.h (#227) 2023-12-19 17:01:59 -08:00
davidkoski
37024d899c
fixes for building with swiftpm (#225)
- clbas is part of veclib (compile failure)
- add SWIFTPM_BUNDLE #define to allow loading the metallib from a swiftpm resource bundle
2023-12-19 16:22:10 -08:00
Angelos Katharopoulos
dfa9f4bc58
An initial quantized matmul implementation (#205)
* Add quantized matvec
* Add quantized matrix matrix with 2nd matrix transposed
* Add quantized matmul tests
* Add a slow cpu quantized matmul
* Add a slightly faster vectorized cpu version
2023-12-18 23:18:57 -08:00
Abe Leininger
e6872a4149
Added linspace (#181)
* linspace ops support

---------

Co-authored-by: Awni Hannun <awni@apple.com>
2023-12-18 19:57:55 -08:00
Angelos Katharopoulos
4d4af12c6f
Adds round op and primitive (#203) 2023-12-18 11:32:48 -08:00
Awni Hannun
0e5807bbcb
include optional (#202) 2023-12-17 22:01:35 -08:00
Cyril Zakka, MD
8eb56beb3a
Added clip function (#159)
* Added clip

* Added Python bindings

* Formatting

* Added cpp tests

* Added Python tests

* python bindings work

* rebase

---------

Co-authored-by: Awni Hannun <awni@apple.com>
2023-12-17 20:00:29 -08:00
Awni Hannun
90d04072b7
fix build w/ flatten (#195) 2023-12-17 11:58:45 -08:00
__mo_san__
52e1589a52
implemented Flatten Module (#149)
* implemented flatten op

---------

Co-authored-by: Awni Hannun <awni@apple.com>
2023-12-16 21:54:37 -08:00
Diogo
dc2edc762c
added tri / tril / triu (#170)
* added tri / tril / triu

* fixed tests

* ctest tests

* tri overload and simplified tests

* changes from comment

* more tests for m

* ensure assert if not 2-D

* remove broadcast_to

* minor tweaks

---------

Co-authored-by: Awni Hannun <awni@apple.com>
2023-12-15 17:30:34 -08:00
Awni Hannun
2e02acdc83
add base kwarg to rope (#186) 2023-12-15 16:47:59 -08:00
Ronan Collobert
83f266c44c
Lazy metal_device_ initialization (#185)
This ensures it is defined when the Scheduler needs it.
2023-12-15 16:06:46 -08:00
Víctor Aguilar
f24200db2c
accross -> across (#183) 2023-12-15 13:46:50 -08:00
Jason
e28b57e371
Added mx.stack c++ frontend impl (#123)
* stack C++ operation + python bindings
2023-12-14 13:21:19 -08:00
Awni Hannun
e5851e52b1
Add move and swap axis, and vmap for slice, concat, and gather (#158)
* add move and swap axis, and vmap for slice, concat, and gather
2023-12-14 12:59:12 -08:00
Luca Arnaboldi
b93c4cf378
Floor and Ceil (#150)
* Implements Floor and Ceil Ops
2023-12-14 10:00:23 -08:00
Ikko Eltociear Ashimine
c3272d4917
Update conv.cpp (#145)
Peform -> Perform
2023-12-12 11:27:49 -08:00
Cyril Zakka, MD
e080290ba4
Added eye/identity ops (#119)
`eye` and `identity` C++ and Python ops
2023-12-11 12:38:17 -08:00
Awni Hannun
71d1fff90a
Bug fix in metal binary kernel dispatch for large arrays (#125)
* bug fix

* format
2023-12-10 16:12:31 -08:00
Angelos Katharopoulos
600db7d754
Fix build on Xcode 14 (#116)
* Fix build on Xcode 14

* Style fixes
2023-12-10 06:58:52 -08:00
Angelos Katharopoulos
2b714714e1
Add the remainder op (#85)
* Add remainder in the C++ backend
* Add the python binding and test
2023-12-08 15:08:52 -08:00
Angelos Katharopoulos
209404239b
Fix the accelerate dispatch for the power op (#70)
- The exponent and base were swapped because accelerate is using
  exponent-base instead of base-exponent
- Fix also the test for binary ops as it was testing op(x, x) which
  couldn't catch ordering errors like that
2023-12-08 10:58:03 -08:00
Awni Hannun
4e3bdb560c
random generation fix (#80)
Random generation fix
2023-12-08 10:40:57 -08:00
Jagrit Digani
d518b3b6a5
Fix gemv broadcasting bug (#6)
* Fix broadcasting bug in gemv
* Add relevant tests in test_blas.py
2023-12-05 14:15:43 -08:00
Awni Hannun
db487e6b1a format 2023-11-30 11:50:50 -08:00
Awni Hannun
46a39e5b1f copyright + ack 2023-11-30 11:12:53 -08:00
Awni Hannun
c1b6bf3f33 missing file 2023-11-29 12:38:32 -08:00
Jagrit Digani
e6306cfee9 jagrit's commit files 2023-11-29 10:52:08 -08:00
Angelos Katharopoulos
d1f86272a2 angelos's commit files 2023-11-29 10:42:59 -08:00
Awni Hannun
8ca7f9e8e9 awni's commit files 2023-11-29 10:30:41 -08:00