fix MLFLOW example (#2575)

Co-authored-by: Yuan-Ting Hsieh (謝沅廷) <yuantingh@nvidia.com>

examples/advanced/experiment-tracking/mlflow/jobs/hello-pt-mlflow/app/config/config_fed_server.conf

@@ -48,6 +48,7 @@
       "id": "mlflow_receiver_with_tracking_uri",
       "path": "nvflare.app_opt.tracking.mlflow.mlflow_receiver.MLflowReceiver",
       "args": {
+        tracking_uri = "file:///{WORKSPACE}/{JOB_ID}/mlruns"
         "kwargs": {
           "experiment_name": "hello-pt-experiment",
           "run_name": "hello-pt-with-mlflow",

examples/hello-world/step-by-step/cifar10/code/fl/train_with_mlflow.py

@@ -139,7 +139,7 @@ def evaluate(input_weights):
                     running_loss += loss.item()
                     if i % 2000 == 1999:  # print every 2000 mini-batches
                         print(f"({client_id}) [{epoch + 1}, {i + 1:5d}] loss: {running_loss / 2000:.3f}")
-                        global_step = input_model.current_round * local_epochs * batch_size + epoch * batch_size + i
+                        global_step = input_model.current_round * steps + epoch * len(trainloader) + i
                         mlflow.log_metric("loss", running_loss / 2000, global_step)
                         running_loss = 0.0
 

examples/hello-world/step-by-step/cifar10/sag/sag.ipynb

@@ -232,8 +232,8 @@
    "source": [
     "! nvflare job create -j /tmp/nvflare/jobs/cifar10_sag_pt -w sag_pt_in_proc \\\n",
     "-f meta.conf min_clients=2 \\\n",
-    "-f config_fed_client.conf app_script=train.py app_config=\"--batch_size 4 --dataset_path {CIFAR10_ROOT} --num_workers 2\" \\\n",
-    "-f config_fed_server.conf num_rounds=5 \\\n",
+    "-f config_fed_client.conf app_script=train_with_mlflow.py app_config=\"--batch_size 4 --dataset_path {CIFAR10_ROOT} --num_workers 2\" \\\n",
+    "-f config_fed_server.conf num_rounds=2 \\\n",
     "-sd ../code/fl \\\n",
     "-force"
    ]
@@ -289,6 +289,48 @@
     "The next 5 examples will use the same ScatterAndGather workflow, but will demonstrate different execution APIs and feature.\n",
     "In the next example [sag_deploy_map](../sag_deploy_map/sag_deploy_map.ipynb), we will learn about the deploy_map configuration for deployment of apps to different sites."
    ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a49b430b-a65b-4b1e-8793-9b3befcfcfd9",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "!tree  /tmp/nvflare/jobs/cifar10_sag_pt_workspace/"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "50594df7-b4c9-4e5e-944a-403b5a105c27",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "!mlflow ui --port 5000 --backend-store-uri /tmp/nvflare/jobs/cifar10_sag_pt_workspace/server/simulate_job/mlruns"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "af2b6628-61af-4bc8-84d4-a9876a27c7c2",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!tensorboard  --logdir=/tmp/nvflare/jobs/cifar10_sag_pt_workspace/server/simulate_job/tb_events"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d3ad11c3-6ef7-46cd-8778-0090505b14e1",
+   "metadata": {},
+   "outputs": [],
+   "source": []
   }
  ],
  "metadata": {

job_templates/sag_pt_in_proc/config_fed_server.conf

@@ -107,7 +107,7 @@
           path = "nvflare.app_opt.tracking.mlflow.mlflow_receiver.MLflowReceiver"
           args {
             # tracking_uri = "http://0.0.0.0:5000"
-            tracking_uri = ""
+            tracking_uri =  "file:///{WORKSPACE}/{JOB_ID}/mlruns"
             kwargs {
               experiment_name =  "nvflare-sag-pt-experiment"
               run_name = "nvflare-sag-pt-with-mlflow"