From: Francois Fleuret <francois@fleuret.org>
Date: Sat, 3 Apr 2021 10:36:18 +0000 (+0200)
Subject: Simplified with Einstein summations.
X-Git-Url: https://fleuret.org/cgi-bin/gitweb/gitweb.cgi?a=commitdiff_plain;h=fdd573490e517d38fb0477ae1b5df12b74718d45;p=pytorch.git

Simplified with Einstein summations.
---

diff --git a/attentiontoy1d.py b/attentiontoy1d.py
index 2cecad8..d389f0c 100755
--- a/attentiontoy1d.py
+++ b/attentiontoy1d.py
@@ -7,7 +7,7 @@
 
 import torch, math, sys, argparse
 
-from torch import nn
+from torch import nn, einsum
 from torch.nn import functional as F
 
 import matplotlib.pyplot as plt
@@ -190,9 +190,9 @@ class AttentionLayer(nn.Module):
         Q = self.conv_Q(x)
         K = self.conv_K(x)
         V = self.conv_V(x)
-        A = Q.permute(0, 2, 1).matmul(K).softmax(2)
-        x = A.matmul(V.permute(0, 2, 1)).permute(0, 2, 1)
-        return x
+        A = einsum('nct,ncs->nts', Q, K).softmax(2)
+        y = einsum('nts,ncs->nct', A, V)
+        return y
 
     def __repr__(self):
         return self._get_name() + \
@@ -205,7 +205,8 @@ class AttentionLayer(nn.Module):
     def attention(self, x):
         Q = self.conv_Q(x)
         K = self.conv_K(x)
-        return Q.permute(0, 2, 1).matmul(K).softmax(2)
+        A = einsum('nct,ncs->nts', Q, K).softmax(2)
+        return A
 
 ######################################################################