473 Update NVFlare example

federated_learning/nvflare/nvflare_spleen_example/3-Admin.ipynb

@@ -27,7 +27,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -46,17 +46,9 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "a working folder for the admin exists!!!\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "admin_path = \"poc/admin/startup/\"\n",
     "\n",
@@ -73,50 +65,28 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
     "!mkdir -p poc/admin/transfer\n",
-    "!cp -r hello_monai/ poc/admin/transfer/"
+    "!cp -r hello-monai/ poc/admin/transfer/"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "['hello_monai']"
-      ]
-     },
-     "execution_count": 4,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
+   "outputs": [],
    "source": [
     "os.listdir(\"poc/admin/transfer/\")"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "['fl_admin.sh']"
-      ]
-     },
-     "execution_count": 5,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
+   "outputs": [],
    "source": [
     "os.listdir(admin_path)"
    ]
@@ -144,23 +114,9 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/html": [
-       "<a href=\"\", data-commandlinker-command=\"terminal:create-new\"> Open a new terminal</a>"
-      ],
-      "text/plain": [
-       "<IPython.core.display.HTML object>"
-      ]
-     },
-     "execution_count": 6,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
+   "outputs": [],
    "source": [
     "HTML('<a href=\"\", data-commandlinker-command=\"terminal:create-new\"> Open a new terminal</a>')"
    ]
@@ -192,51 +148,18 @@
     "\n",
     "The commands can be:\n",
     "```\n",
-    "upload_app hello_monai\n",
+    "upload_app hello-monai\n",
     "set_run_number 1\n",
-    "deploy_app hello_monai server\n",
-    "deploy_app hello_monai client\n",
-    "```\n",
-    "\n",
-    "Now, let's check if the folder has been distributed into the server and all client(s):"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "config files on server: ['app_server', 'fl_app.txt']\n",
-      " \n",
-      "config files on site-1: ['app_site-1', 'fl_app.txt']\n",
-      " \n",
-      "config files on site-2: ['fl_app.txt', 'app_site-2']\n",
-      " \n"
-     ]
-    }
-   ],
-   "source": [
-    "run_file = \"run_1\"\n",
-    "\n",
-    "poc_path = \"poc/\"\n",
-    "\n",
-    "for name in [\"server\", \"site-1\", \"site-2\"]:\n",
-    "    path = os.path.join(poc_path, name, run_file)\n",
-    "    if os.path.exists(path):\n",
-    "        print(\"config files on {}: {}\".\n",
-    "              format(name, os.listdir(path)))\n",
-    "        print(\" \")"
+    "deploy_app hello-monai server\n",
+    "deploy_app hello-monai client\n",
+    "```"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "This example prepares two different data list files: `dataset_part1.json` and `dataset_part2.json`, and they have the same validation set and totally different training set. The default file used in `config_train.json` is `config/dataset_part1.json`. Therefore, if you want to let two clients train on different data, you can switch to use `dataset_part2.json` for `org1-b`.\n",
+    "This example prepares two different data list files: `dataset_part1.json` and `dataset_part2.json`, and they have the same validation set and totally different training sets. The default file used in `config_train.json` is `config/dataset_part1.json`. Therefore, if you want to let two clients train on different data, you can switch to use `dataset_part2.json` for `org1-b`.\n",
     "\n",
     "[Link to site-1 config](poc/site-1/run_1/app_site-1/config/config_train.json)\n",
     "\n",
@@ -250,13 +173,12 @@
     "### (Optional) Copy Dataset\n",
     "\n",
     "After starting a client (for example `site-1`), the Spleen dataset will be downloaded into:\n",
-    "`run_1/app_site-1/Task09_Spleen.tar`.\n",
+    "`poc/site-1/`.\n",
     "\n",
-    "To prevent repeatedly downloading the dataset, you can copy the uncompressed `Task09_Spleen` into the corresponding place after running the `deploy_app` command.\n",
-    "For example:\n",
+    "If you already have the `Task09_Spleen`, you can copy it into the corresponding place directly to prevent repeatedly download. For example:\n",
     "\n",
     "```\n",
-    "cp -r /path-to-dataset/Task09_Spleen poc/site-1/run_2/app_site-1/\n",
+    "cp -r /path-to-dataset/Task09_Spleen poc/site-1/\n",
     "```"
    ]
   },

federated_learning/nvflare/nvflare_spleen_example/README.md

@@ -2,13 +2,15 @@
 
 ## Brief Introduction
 
-This repository contains an end-to-end Federated training example based on MONAI trainers and [NVIDIA FLARE](https://github.com/nvidia/nvflare).
+This repository contains an end-to-end Federated training example based on MONAI trainers and [NVIDIA FLARE](https://github.com/nvidia/nvflare). Please also download the `hello-monai` folder in [NVFlare/examples](https://github.com/NVIDIA/NVFlare/tree/main/examples/hello-monai), and copy it into this directory.
+
+
 
 This example requires Python 3.8.10
 
-Inside this folder:
+Inside this directory:
 - All Jupiter notebooks are used to build an FL experiment step-by-step.
-- `hello-monai` is a folder containing all required config files for the experiment (in `config/`) and the customized trainer (in `custom`) and its components.
+- hello-monai (needs to be downloaded) is a folder containing all required config files for the experiment (in `config/`) and the customized trainer (in `custom`) and its components.
 
 ## Installation