homework5

homework5/transformer_nmt_student.ipynb

@@ -1038,7 +1038,7 @@
     "        # Self-Attention：注意self-attention的q，k和v均为decoder hidden\n",
     "        x = self.sublayer[0](x, lambda x: self.self_attn(x, x, x, tgt_mask))\n",
     "        # Context-Attention：注意context-attention的q为decoder hidden，而k和v为encoder hidden\n",
-    "        x = self.sublayer[1](x, lambda x: self.self_attn(x, m, m, src_mask))\n",
+    "        x = self.sublayer[1](x, lambda x: self.src_attn(x, m, m, src_mask))\n",
     "        return self.sublayer[2](x, self.feed_forward)"
    ]
   },

homework5/transformer_nmt_student.py

@@ -742,7 +742,7 @@ class DecoderLayer(nn.Module):
         # Self-Attention：注意self-attention的q，k和v均为decoder hidden
         x = self.sublayer[0](x, lambda x: self.self_attn(x, x, x, tgt_mask))
         # Context-Attention：注意context-attention的q为decoder hidden，而k和v为encoder hidden
-        x = self.sublayer[1](x, lambda x: self.self_attn(x, m, m, src_mask))
+        x = self.sublayer[1](x, lambda x: self.src_attn(x, m, m, src_mask))
         return self.sublayer[2](x, self.feed_forward)