Merge pull request #183 from superkostya/fix-mha-wrapper-class

Fix an incorrect input dimension

ch03/01_main-chapter-code/multihead-attention.ipynb

@@ -38,7 +38,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "torch version: 2.2.1\n"
+      "torch version: 2.2.2\n"
      ]
     }
    ],
@@ -365,6 +365,14 @@
     "\n",
     "print(\"context_vecs.shape:\", context_vecs.shape)"
    ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f1d965a5-9b98-4554-8646-7ecd497874cb",
+   "metadata": {},
+   "outputs": [],
+   "source": []
   }
  ],
  "metadata": {

ch03/02_bonus_efficient-multihead-attention/mha-implementations.ipynb

@@ -341,7 +341,7 @@
     "        self.d_out = d_out\n",
     "\n",
     "        self.qkv = nn.Linear(d_in, 3 * d_out, bias=qkv_bias)\n",
-    "        self.proj = nn.Linear(d_in, d_out)\n",
+    "        self.proj = nn.Linear(d_out, d_out)\n",
     "        self.dropout = dropout\n",
     "\n",
     "    def forward(self, x):\n",