x-transformers 1.30.9__py3-none-any.whl → 1.30.11__py3-none-any.whl

x_transformers/x_transformers.py

@@ -311,19 +311,33 @@ class CoPE(Module):
     def __init__ (
         self,
         dim,
+        heads,
         max_pos,
         soft_onehot = False,
+        talking_heads = False,
+        soft_onehot_temp = 5e-2
     ):
         super () . __init__ ()
         self.max_pos = max_pos
         self.pos_emb = nn.Parameter(torch.zeros(max_pos, dim))
 
+        self.talking_heads = nn.Conv2d(heads, heads, 1, bias = False) if talking_heads else None
         self.soft_onehot = soft_onehot
+        self.soft_onehot_temp = soft_onehot_temp
 
         if soft_onehot:
             self.register_buffer('positions', torch.arange(max_pos))
 
-    def forward(self, query, attn_logits, temp = 5e-2):
+    def forward(self, query, attn_logits):
+
+        if exists(self.talking_heads):
+            i, j = attn_logits.shape[-2:]
+            causal_mask = attn_logits.new_ones(i, j).triu_(j - i + 1).bool()
+
+            attn_logits = self.talking_heads(attn_logits)
+
+            attn_logits = attn_logits.masked_fill(causal_mask, -torch.finfo(attn_logits.dtype).max)
+
         # compute positions
 
         gates = attn_logits.sigmoid()
@@ -335,7 +349,7 @@ class CoPE(Module):
 
         if self.soft_onehot:
             diff_pos = (pos[..., None] - self.positions).abs()
-            soft_onehot_pos = F.softmax(-diff_pos / temp, dim = -1)
+            soft_onehot_pos = F.softmax(-diff_pos / self.soft_onehot_temp, dim = -1)
             cope_pos_emb = einsum('b h i j p, b h i p -> b h i j', soft_onehot_pos, logits_int)
         else:
             # interpolate from integer positions
@@ -770,6 +784,7 @@ class Attention(Module):
         use_cope = False,
         cope_max_pos = 16,
         cope_soft_onehot_pos = False,
+        cope_talking_heads = False,
         logit_softclamp_value = None,
         onnxable = False
     ):
@@ -853,7 +868,13 @@ class Attention(Module):
             assert causal, 'CoPE was designed for causal attention'
             assert not flash, 'CoPE is not flash attention compatible'
 
-            cope = CoPE(dim_head, cope_max_pos, soft_onehot = cope_soft_onehot_pos)
+            cope = CoPE(
+                dim = dim_head,
+                heads = heads,
+                max_pos = cope_max_pos,
+                talking_heads = cope_talking_heads,
+                soft_onehot = cope_soft_onehot_pos
+            )
 
         # attend class - includes core attention algorithm + talking heads
 

{x_transformers-1.30.9.dist-info → x_transformers-1.30.11.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: x-transformers
-Version: 1.30.9
+Version: 1.30.11
 Summary: X-Transformers - Pytorch
 Home-page: https://github.com/lucidrains/x-transformers
 Author: Phil Wang

{x_transformers-1.30.9.dist-info → x_transformers-1.30.11.dist-info}/RECORD

@@ -4,11 +4,11 @@ x_transformers/autoregressive_wrapper.py,sha256=uX8Mb0zLsQrZECt_9UGt35g7tC05Rk3n
 x_transformers/continuous.py,sha256=WO52n9lFAXv5-SGadi2cApGF8dkouN8QSTEOuC7erj8,6180
 x_transformers/dpo.py,sha256=LjvWgCkqTl-UuehrzQ8nkX5guLr4whYwsmm7SKSwdls,3450
 x_transformers/nonautoregressive_wrapper.py,sha256=ys_p8obc7lTeeodCqvkRKxOXQ1C9T3j5Jwr-JbVgnXk,10432
-x_transformers/x_transformers.py,sha256=V_kR77GhWNu3TuUetu3xryWoFliwTJdNKRX2lVWnsRc,68274
+x_transformers/x_transformers.py,sha256=97WREYHfGljhrveGNUUSEWk2xbFvKdM52QXW7cnoBpk,69019
 x_transformers/xl_autoregressive_wrapper.py,sha256=DCx4n0_c1tFai4nOqaWVnqx2p9eutsZsDMiMP1ckxNU,4117
 x_transformers/xval.py,sha256=QE1ltYZTR_eGgIHPP2BrMWVWVLqMW-OpDZh87BSmQEg,8563
-x_transformers-1.30.9.dist-info/LICENSE,sha256=As9u198X-U-vph5noInuUfqsAG2zX_oXPHDmdjwlPPY,1066
-x_transformers-1.30.9.dist-info/METADATA,sha256=auoB6F1DCe054gY70tE1k048x6wQRtftk8-Pk6nJD-I,661
-x_transformers-1.30.9.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-x_transformers-1.30.9.dist-info/top_level.txt,sha256=hO6KGpFuGucRNEtRfme4A_rGcM53AKwGP7RVlRIxS5Q,15
-x_transformers-1.30.9.dist-info/RECORD,,
+x_transformers-1.30.11.dist-info/LICENSE,sha256=As9u198X-U-vph5noInuUfqsAG2zX_oXPHDmdjwlPPY,1066
+x_transformers-1.30.11.dist-info/METADATA,sha256=kpv8_mEb4DU4zL_oyLwzbWJ7z8WV7mf7c466Hehn-6c,662
+x_transformers-1.30.11.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+x_transformers-1.30.11.dist-info/top_level.txt,sha256=hO6KGpFuGucRNEtRfme4A_rGcM53AKwGP7RVlRIxS5Q,15
+x_transformers-1.30.11.dist-info/RECORD,,