x-transformers 2.5.2__tar.gz → 2.5.3__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.
Files changed (63) hide show
  1. {x_transformers-2.5.2 → x_transformers-2.5.3}/PKG-INFO +1 -1
  2. {x_transformers-2.5.2 → x_transformers-2.5.3}/pyproject.toml +1 -1
  3. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/x_transformers.py +4 -0
  4. {x_transformers-2.5.2 → x_transformers-2.5.3}/.github/FUNDING.yml +0 -0
  5. {x_transformers-2.5.2 → x_transformers-2.5.3}/.github/workflows/python-publish.yml +0 -0
  6. {x_transformers-2.5.2 → x_transformers-2.5.3}/.github/workflows/python-test.yaml +0 -0
  7. {x_transformers-2.5.2 → x_transformers-2.5.3}/.gitignore +0 -0
  8. {x_transformers-2.5.2 → x_transformers-2.5.3}/LICENSE +0 -0
  9. {x_transformers-2.5.2 → x_transformers-2.5.3}/README.md +0 -0
  10. {x_transformers-2.5.2 → x_transformers-2.5.3}/data/README.md +0 -0
  11. {x_transformers-2.5.2 → x_transformers-2.5.3}/data/enwik8.gz +0 -0
  12. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/all-attention.png +0 -0
  13. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/attention-on-attention.png +0 -0
  14. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/cosine-sim-attention.png +0 -0
  15. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/deepnorm.png +0 -0
  16. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/dynamic-pos-bias-linear.png +0 -0
  17. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/dynamic-pos-bias-log.png +0 -0
  18. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/dynamic-pos-bias-sinusoidal.png +0 -0
  19. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/dynamic-pos-bias.png +0 -0
  20. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/enhanced-recurrence.png +0 -0
  21. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/fcm.png +0 -0
  22. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/ffglu.png +0 -0
  23. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/flash-attention.png +0 -0
  24. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/gate_values.png +0 -0
  25. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/gating.png +0 -0
  26. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/length-extrapolation-scale.png +0 -0
  27. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/macaron-1.png +0 -0
  28. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/macaron-2.png +0 -0
  29. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/memory-transformer.png +0 -0
  30. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/normformer.png +0 -0
  31. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/pia.png +0 -0
  32. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/qknorm-analysis.png +0 -0
  33. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/resi_dual.png +0 -0
  34. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/residual_attn.png +0 -0
  35. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/rezero.png +0 -0
  36. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/rotary.png +0 -0
  37. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/sandwich-2.png +0 -0
  38. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/sandwich.png +0 -0
  39. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/sandwich_norm.png +0 -0
  40. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/scalenorm.png +0 -0
  41. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/talking-heads.png +0 -0
  42. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/topk-attention.png +0 -0
  43. {x_transformers-2.5.2 → x_transformers-2.5.3}/images/xval.png +0 -0
  44. {x_transformers-2.5.2 → x_transformers-2.5.3}/tests/test_x_transformers.py +0 -0
  45. {x_transformers-2.5.2 → x_transformers-2.5.3}/train_belief_state.py +0 -0
  46. {x_transformers-2.5.2 → x_transformers-2.5.3}/train_copy.py +0 -0
  47. {x_transformers-2.5.2 → x_transformers-2.5.3}/train_entropy_tokenizer.py +0 -0
  48. {x_transformers-2.5.2 → x_transformers-2.5.3}/train_enwik8.py +0 -0
  49. {x_transformers-2.5.2 → x_transformers-2.5.3}/train_length_extrapolate.py +0 -0
  50. {x_transformers-2.5.2 → x_transformers-2.5.3}/train_parity.py +0 -0
  51. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/__init__.py +0 -0
  52. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/attend.py +0 -0
  53. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/autoregressive_wrapper.py +0 -0
  54. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/belief_state_wrapper.py +0 -0
  55. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/continuous.py +0 -0
  56. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/dpo.py +0 -0
  57. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/entropy_based_tokenizer.py +0 -0
  58. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/multi_input.py +0 -0
  59. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/neo_mlp.py +0 -0
  60. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/nonautoregressive_wrapper.py +0 -0
  61. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/up_wrapper.py +0 -0
  62. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/xl_autoregressive_wrapper.py +0 -0
  63. {x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/xval.py +0 -0
{x_transformers-2.5.2 → x_transformers-2.5.3}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: x-transformers
3
- Version: 2.5.2
3
+ Version: 2.5.3
4
4
  Summary: X-Transformers
5
5
  Project-URL: Homepage, https://pypi.org/project/x-transformers/
6
6
  Project-URL: Repository, https://github.com/lucidrains/x-transformers
{x_transformers-2.5.2 → x_transformers-2.5.3}/pyproject.toml RENAMED
@@ -1,6 +1,6 @@
1
1
  [project]
2
2
  name = "x-transformers"
3
- version = "2.5.2"
3
+ version = "2.5.3"
4
4
  description = "X-Transformers"
5
5
  authors = [
6
6
  { name = "Phil Wang", email = "lucidrains@gmail.com" }
{x_transformers-2.5.2 → x_transformers-2.5.3}/x_transformers/x_transformers.py RENAMED
@@ -2787,6 +2787,7 @@ class AttentionPool(Module):
2787
2787
  self.pooler = Attention(dim = dim, dim_context = dim_context, heads = heads, dim_head = dim_head, **attn_kwargs)
2788
2788
 
2789
2789
  self.add_residual = add_residual
2790
+ self.squeeze_output = squeeze_output
2790
2791
 
2791
2792
  def forward(self, context, mask = None):
2792
2793
  batch = context.shape[0]
@@ -2798,6 +2799,9 @@ class AttentionPool(Module):
2798
2799
  if self.add_residual:
2799
2800
  pooled = pooled + queries
2800
2801
 
2802
+ if self.squeeze_output:
2803
+ pooled = rearrange(pooled, 'b 1 d -> b d')
2804
+
2801
2805
  return pooled
2802
2806
 
2803
2807
  class ViTransformerWrapper(Module):
{x_transformers-2.5.2 → x_transformers-2.5.3}/LICENSE RENAMED
File without changes
{x_transformers-2.5.2 → x_transformers-2.5.3}/README.md RENAMED
File without changes