mps-flash-attn 0.1.5__tar.gz → 0.1.6__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.

This version of mps-flash-attn might be problematic. Click here for more details.

Files changed (38) hide show
  1. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/PKG-INFO +1 -1
  2. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/__init__.py +5 -2
  3. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn.egg-info/PKG-INFO +1 -1
  4. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/pyproject.toml +1 -1
  5. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/LICENSE +0 -0
  6. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/README.md +0 -0
  7. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/csrc/mps_flash_attn.mm +0 -0
  8. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/06c421e7a01418cf64aafa07f6b1df0558148583959c596d9a7ce260987f89f0.metallib +0 -0
  9. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/09b9615289be632fdf05444004a0b3b67fb1b70b05a7e0fce8e0ba3a95e3921c.metallib +0 -0
  10. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/0c36461301fb52cbad786d0642b020ad2bfc7229b487ccb5dff44d198423b347.metallib +0 -0
  11. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/2ca9312d1151f792e1a95617db9186928300e3d0ffbe016f0ad53b62ab840bac.metallib +0 -0
  12. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/73254c55475c6b7f7b009f095398994b1f9ae8215beafcf810f100357ccc99b2.metallib +0 -0
  13. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/73254c55475c6b7f7b009f095398994b1f9ae8215beafcf810f100357ccc99b2_1024_1024.bin +0 -0
  14. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/73254c55475c6b7f7b009f095398994b1f9ae8215beafcf810f100357ccc99b2_2048_2048.bin +0 -0
  15. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/73254c55475c6b7f7b009f095398994b1f9ae8215beafcf810f100357ccc99b2_4096_4096.bin +0 -0
  16. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/73254c55475c6b7f7b009f095398994b1f9ae8215beafcf810f100357ccc99b2_512_512.bin +0 -0
  17. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/73254c55475c6b7f7b009f095398994b1f9ae8215beafcf810f100357ccc99b2_8192_8192.bin +0 -0
  18. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/771935bf47d248650e287da82bc82e04bff7c4c52964823e7a12462ccd23408e.metallib +0 -0
  19. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/975aece2b4d3d78035be08a0735a7deacf2e544adee5af81c9c0a3a42a926129.metallib +0 -0
  20. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/a5e2c5c401e3872af0899c1fb3e30b5f52a6070fc49c9dac02982cc1c2f25849.metallib +0 -0
  21. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/ac4573fb201e92310867c59bd569a8ae68f859d60a9352d9d4d5d41c1547c83c.metallib +0 -0
  22. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/adc0f77ff05156bbda8fe78afd9ba8a8d3c890ba8fea0902ae79a6ae8c4f04c3.metallib +0 -0
  23. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/eab4f40de4b0ebd2765b41c25dba7ccab5db4abf6a6eb87d76fff7b5e0829b2f.metallib +0 -0
  24. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/eab4f40de4b0ebd2765b41c25dba7ccab5db4abf6a6eb87d76fff7b5e0829b2f_1024_1024.bin +0 -0
  25. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/eab4f40de4b0ebd2765b41c25dba7ccab5db4abf6a6eb87d76fff7b5e0829b2f_2048_2048.bin +0 -0
  26. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/eab4f40de4b0ebd2765b41c25dba7ccab5db4abf6a6eb87d76fff7b5e0829b2f_4096_4096.bin +0 -0
  27. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/eab4f40de4b0ebd2765b41c25dba7ccab5db4abf6a6eb87d76fff7b5e0829b2f_512_512.bin +0 -0
  28. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/eab4f40de4b0ebd2765b41c25dba7ccab5db4abf6a6eb87d76fff7b5e0829b2f_8192_8192.bin +0 -0
  29. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/f08fe0efd72e055177e068154dae01e08c4d52d3cb883330a04f1431d274aece.metallib +0 -0
  30. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/kernels/manifest.json +0 -0
  31. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/lib/libMFABridge.dylib +0 -0
  32. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn.egg-info/SOURCES.txt +0 -0
  33. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn.egg-info/dependency_links.txt +0 -0
  34. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn.egg-info/requires.txt +0 -0
  35. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn.egg-info/top_level.txt +0 -0
  36. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/setup.cfg +0 -0
  37. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/setup.py +0 -0
  38. {mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/tests/test_attention.py +0 -0
{mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: mps-flash-attn
3
- Version: 0.1.5
3
+ Version: 0.1.6
4
4
  Summary: Flash Attention for PyTorch on Apple Silicon (M1/M2/M3/M4)
5
5
  Author: imperatormk
6
6
  License-Expression: MIT
{mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn/__init__.py RENAMED
@@ -4,7 +4,7 @@ MPS Flash Attention - Flash Attention for PyTorch on Apple Silicon
4
4
  This package provides memory-efficient attention using Metal Flash Attention kernels.
5
5
  """
6
6
 
7
- __version__ = "0.1.5"
7
+ __version__ = "0.1.6"
8
8
 
9
9
  import torch
10
10
  from typing import Optional
@@ -174,10 +174,13 @@ def replace_sdpa():
174
174
  def patched_sdpa(query, key, value, attn_mask=None, dropout_p=0.0,
175
175
  is_causal=False, scale=None):
176
176
  # Use MFA for MPS tensors without mask/dropout
177
+ # Only use MFA for seq_len >= 1024 where it outperforms PyTorch's math backend
178
+ # For shorter sequences, PyTorch's simpler matmul+softmax approach is faster
177
179
  if (query.device.type == 'mps' and
178
180
  attn_mask is None and
179
181
  dropout_p == 0.0 and
180
- _HAS_MFA):
182
+ _HAS_MFA and
183
+ query.shape[2] >= 1024):
181
184
  try:
182
185
  return flash_attention(query, key, value, is_causal=is_causal, scale=scale)
183
186
  except Exception:
{mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/mps_flash_attn.egg-info/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: mps-flash-attn
3
- Version: 0.1.5
3
+ Version: 0.1.6
4
4
  Summary: Flash Attention for PyTorch on Apple Silicon (M1/M2/M3/M4)
5
5
  Author: imperatormk
6
6
  License-Expression: MIT
{mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/pyproject.toml RENAMED
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
4
4
 
5
5
  [project]
6
6
  name = "mps-flash-attn"
7
- version = "0.1.5"
7
+ version = "0.1.6"
8
8
  description = "Flash Attention for PyTorch on Apple Silicon (M1/M2/M3/M4)"
9
9
  readme = "README.md"
10
10
  license = "MIT"
{mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/LICENSE RENAMED
File without changes
{mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/README.md RENAMED
File without changes
{mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/setup.cfg RENAMED
File without changes
{mps_flash_attn-0.1.5 → mps_flash_attn-0.1.6}/setup.py RENAMED
File without changes