Awni Hannun
fa89f0b150
faster gather qmm sorted test ( #2463 )
2025-08-05 06:27:40 -07:00
Awni Hannun
5597fa089c
Fix qvm splitk ( #2415 )
2025-07-25 11:50:24 -07:00
Angelos Katharopoulos
4a9b29a875
MoE backward improvements ( #2335 )
2025-07-07 17:59:53 -07:00
Awni Hannun
4fda5fbdf9
add python testing for cuda with ability to skip list of tests ( #2295 )
2025-06-15 10:56:48 -07:00
Awni Hannun
6ef2f67e7f
5bit quants ( #2226 )
...
* 5bit quants
* 5bit quants
2025-05-30 12:12:10 -07:00
Awni Hannun
7bb063bcb3
Enable vjp for quantized scale and bias ( #2129 )
...
* Enable vjp for quantized scale and bias
* higher tol
2025-04-29 13:03:09 -07:00
Angelos Katharopoulos
5de6d94a90
Gather qmm batched kernel and refactoring of quantized ( #2078 )
2025-04-17 13:53:11 -07:00
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
Awni Hannun
0c259961ac
matmul jvps ( #1772 )
2025-01-17 10:36:26 -08:00
Alex Barron
c7b0300af5
Fix batched qmv bug ( #1758 )
2025-01-09 11:45:57 -08:00
Alex Barron
c79f6a4a8c
3 and 6 bit quantization ( #1613 )
...
* Support 3 and 6 bit quantization
2024-11-22 10:22:13 -08:00
Alex Barron
26be608470
Add split_k qvm for long context ( #1564 )
...
* Add splitk qvm
* configurable splitk
* tuning
* remove extra instantiation
* remove refactor
* separate test
* cpu tolerance
2024-11-05 11:25:19 -08:00
Alex Barron
d15fa13daf
Batched Quantized Matmul + Fast Small QMV ( #1503 )
...
* add fast qmv for small dims
* fix test
* batched cpu
* add batched template param
* refactor metal quantized.cpp
2024-10-21 16:23:17 -07:00
Alex Barron
c52d1600f0
Fused Affine Quantize/Dequantize ops ( #1282 )
...
* Add fast affine dequantize
* add full quantize kernel
* fused kernel with scale/bias computation
* fix docstring
* fix no jit error
* fix test
* test fix
* reduce fast api to only affine_quantize
2024-07-29 15:11:38 -07:00
Awni Hannun
d568c7ee36
Rename block sparse ( #1149 )
...
* block_sparse_mm to gather_mm
* rename
* nit
* nit
2024-05-22 07:48:34 -07:00
Angelos Katharopoulos
e78a6518fa
Block sparse qmm ( #1124 )
2024-05-16 15:24:14 -07:00
Angelos Katharopoulos
17f57df797
Improvements in the quantizer and dequantization kernel ( #1061 )
2024-05-01 18:19:11 -07:00
Angelos Katharopoulos
8db7161c94
Bug fix in quantize ( #1054 )
2024-04-29 20:55:04 -07:00
Angelos Katharopoulos
ec8578d41a
Fix quantization of all 0s ( #1028 )
2024-04-24 00:40:42 -07:00
Angelos Katharopoulos
84d61d27aa
Make sure 0 is represented in the quantization ( #1016 )
2024-04-19 19:47:26 -07:00
Awni Hannun
039da779d1
No quant reshape ( #957 )
...
* precise option on cpu
* remove print
* remove reshape in quant matmul
* no quant reshape
2024-04-04 11:52:12 -07:00
Angelos Katharopoulos
5f9ba3019f
Fix qmm_t for unaligned cases ( #923 )
2024-03-28 15:34:57 -07:00
Angelos Katharopoulos
40c108766b
Quantized matmul fix ( #677 )
...
* Fix qmv for small or unaligned matrices
* Fix qmm
2024-02-12 18:54:21 -08:00
Awni Hannun
7a34e46677
Quantize with groups of 32 ( #511 )
...
* allow quantize with group sizes of 32
* missing cpu dispatch
* remove print
* Fix qvm for group_size 32
---------
Co-authored-by: Angelos Katharopoulos <a_katharopoulos@apple.com >
2024-01-21 06:19:05 -08:00
Angelos Katharopoulos
c15fe3e61b
Allow arbitrary first dimension in quantization kernels. ( #458 )
...
* Allow arbitrary first dim on qmm_t and qmv
* Allow arbitrary first dim on qmm and qvm
* Specialized aligned vs unaligned case
* Add more checks for valid quantizations
2024-01-16 00:46:21 -08:00
Angelos Katharopoulos
e7f5059fe4
Support for quantized matmul with w and w^T ( #349 )
...
* Add the metal qvm implementation
* Add qmm_n
* Add gradient wrt to input for quantized_matmul
2024-01-03 14:22:36 -08:00
Angelos Katharopoulos
447bc089b9
Fix tolerance in de-/quantization test ( #295 )
2023-12-26 19:21:05 -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
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
