Awni Hannun 4758c8baa1 Start to cleanup/unify accelerate and common back-ends (Part 1/N) (#1777) ...

* start to cleanup/unify accelerate and common back-ends

* more progress

* simplify

* add half type and allow infs in simd exp

* unify softmax + quantized, more dispatches to simd quantized mm

* add sin/cos, use simd in vector-scalar ops

* faster CPU vectorize quant

* faster erf/erfinv

2025-01-29 14:34:49 -08:00

..

3rdparty

jagrit's commit files

2023-11-29 10:52:08 -08:00

backend

Start to cleanup/unify accelerate and common back-ends (Part 1/N) (#1777)

2025-01-29 14:34:49 -08:00

distributed

Ring distributed backend (#1784)

2025-01-27 22:15:01 -08:00

io

Ring distributed backend (#1784)

2025-01-27 22:15:01 -08:00

types

Start to cleanup/unify accelerate and common back-ends (Part 1/N) (#1777)

2025-01-29 14:34:49 -08:00

allocator.cpp

Fix sibling leak (#1590)

2024-11-18 19:17:01 -08:00

allocator.h

Allow querying the allocator for the buffer size (#1404)

2024-09-11 21:02:16 -07:00

array.cpp

Dynamic broadcasting for shapeless compile/export (#1722)

2025-01-09 11:04:24 -08:00

array.h

Dynamic slicing (#1741)

2025-01-07 14:02:16 -08:00

CMakeLists.txt

Export / import functions to / from a file (#1642)

2024-12-24 11:19:13 -08:00

compile_impl.h

Simplify removes no-ops from the tape (#1759)

2025-01-09 11:23:19 -08:00

compile.cpp

recompile when shapeless is different (#1776)

2025-01-20 21:07:10 -08:00

compile.h

allow compiling lambdas in C++ (#1650)

2024-12-06 13:13:21 -08:00

device.cpp

copyright + ack

2023-11-30 11:12:53 -08:00

device.h

minor fixes (#1194)

2024-06-12 22:06:49 -07:00

dtype.cpp

dtype is copy assignable (#1436)

2024-09-25 12:07:13 -07:00

dtype.h

dtype is copy assignable (#1436)

2024-09-25 12:07:13 -07:00

einsum.cpp

Einsum ellipsis (#1788)

2025-01-25 01:28:03 -08:00

einsum.h

Einsum (#1269)

2024-07-25 09:36:44 -07:00

event.h

Fix array is_available race cases (#1468)

2024-10-07 19:13:50 -07:00

export_impl.h

Export / import functions to / from a file (#1642)

2024-12-24 11:19:13 -08:00

export.cpp

Dynamic broadcasting for shapeless compile/export (#1722)

2025-01-09 11:04:24 -08:00

export.h

Export / import functions to / from a file (#1642)

2024-12-24 11:19:13 -08:00

fast_primitives.h

Export / import functions to / from a file (#1642)

2024-12-24 11:19:13 -08:00

fast.cpp

Fix rope fallback to not upcast (#1797)

2025-01-26 19:07:21 -08:00

fast.h

Allow offset to be an mx.array for mx.fast.rope (#1724)

2024-12-19 15:51:44 -08:00

fft.cpp

Use int64 stride everywhere (#1671)

2024-12-09 11:09:02 -08:00

fft.h

Use int64 stride everywhere (#1671)

2024-12-09 11:09:02 -08:00

graph_utils.cpp

Optionally specify names for arrays when exporting (#1749)

2025-01-06 13:07:46 -08:00

graph_utils.h

Optionally specify names for arrays when exporting (#1749)

2025-01-06 13:07:46 -08:00

io.h

Added missing unordered_map includes (#1635)

2024-12-02 07:03:03 -08:00

linalg.cpp

catch stream errors earlier to avoid aborts (#1801)

2025-01-27 14:05:43 -08:00

linalg.h

Eigenvalues and eigenvectors (#1334)

2024-10-22 12:18:48 -07:00

mlx.h

Export / import functions to / from a file (#1642)

2024-12-24 11:19:13 -08:00

ops.cpp

add dilation for conv 3d layers + test for 3d conv w/ dilation (#1802)

2025-01-28 06:17:07 -08:00

ops.h

Ring distributed backend (#1784)

2025-01-27 22:15:01 -08:00

primitives.cpp

matmul jvps (#1772)

2025-01-17 10:36:26 -08:00

primitives.h

Start to cleanup/unify accelerate and common back-ends (Part 1/N) (#1777)

2025-01-29 14:34:49 -08:00

random.cpp

Use int64 stride everywhere (#1671)

2024-12-09 11:09:02 -08:00

random.h

Use int64 stride everywhere (#1671)

2024-12-09 11:09:02 -08:00

scheduler.cpp

Export / import functions to / from a file (#1642)

2024-12-24 11:19:13 -08:00

scheduler.h

Export / import functions to / from a file (#1642)

2024-12-24 11:19:13 -08:00

stream.h

Export / import functions to / from a file (#1642)

2024-12-24 11:19:13 -08:00

threadpool.h

Ring distributed backend (#1784)

2025-01-27 22:15:01 -08:00

transforms_impl.h

fix shapeless compile on ubuntu24 (#1775)

2025-01-18 06:04:36 -08:00

transforms.cpp

fix shapeless compile on ubuntu24 (#1775)

2025-01-18 06:04:36 -08:00

transforms.h

Export / import functions to / from a file (#1642)

2024-12-24 11:19:13 -08:00

utils.cpp

Dynamic broadcasting for shapeless compile/export (#1722)

2025-01-09 11:04:24 -08:00

utils.h

Faster synchronization Fence primitive (#1773)

2025-01-17 18:42:19 -08:00