update hapi and image_segmentation

paddle2.0_docs/high_level_api/high_level_api.ipynb

@@ -384,15 +384,23 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# 场景1：动态图模式\n",
-    "\n",
     "# 使用GPU训练\n",
     "paddle.set_device('gpu')\n",
+    "\n",
     "# 模型封装\n",
-    "model = paddle.Model(mnist)\n",
     "\n",
+    "## 场景1：动态图模式\n",
+    "## 1.1 为模型预测部署场景进行模型训练\n",
+    "## 需要添加input和label数据描述，否则会导致使用model.save(training=False)保存的预测模型在使用时出错\n",
+    "input = paddle.static.InputSpec([None, 1, 28, 28], dtype='float32')\n",
+    "label = paddle.static.InputSpec([None, 1], dtype='int8')\n",
+    "model = paddle.Model(mnist, input, label)\n",
+    "\n",
+    "## 1.2 面向实验而进行的模型训练\n",
+    "## 可以不传递input和label信息\n",
+    "# model = paddle.Model(mnist)\n",
     "\n",
-    "# 场景2：静态图模式\n",
+    "## 场景2：静态图模式\n",
     "# paddle.enable_static()\n",
     "# paddle.set_device('gpu')\n",
     "# input = paddle.static.InputSpec([None, 1, 28, 28], dtype='float32')\n",
@@ -584,6 +592,21 @@
     "          verbose=1)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "#### 注：\n",
+    "\n",
+    "`fit()`的第一个参数不仅可以传递数据集`paddle.io.Dataset`，还可以传递DataLoader，如果想要实现某个自定义的数据集抽样等逻辑，可以在fit外自定义DataLoader，然后传递给fit函数。\n",
+    "\n",
+    "```python\n",
+    "train_dataloader = paddle.io.DataLoader(train_dataset)\n",
+    "...\n",
+    "model.fit(train_dataloader, ...)\n",
+    "```"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},