update hello-flower examples

test tb streaming

reformatting

udpate readme

reformatting

Add client api DLI notebook (NVIDIA#3220)

Add client api DLI notebook.

Add client api DLI notebook.

<!--- Put an `x` in all the boxes that apply, and remove the not
applicable items -->
- [x] Non-breaking change (fix or new feature that would not break
existing functionality).
- [ ] Breaking change (fix or new feature that would cause existing
functionality to change).
- [ ] New tests added to cover the changes.
- [ ] Quick tests passed locally by running `./runtest.sh`.
- [ ] In-line docstrings updated.
- [ ] Documentation updated.

examples/hello-world/hello-flower/README.md

@@ -23,7 +23,10 @@ If you haven't already, we recommend creating a virtual environment.
 python3 -m venv nvflare_flwr
 source nvflare_flwr/bin/activate
 ```
-
+We recommend installing an older version of NumPy as torch/torchvision doesn't support NumPy 2 at this time.
+```bash
+pip install numpy==1.26.4
+```
 ## 2.1 Run a simulation
 
 To run flwr-pt job with NVFlare, we first need to install its dependencies.
@@ -49,3 +52,23 @@ the TensorBoard metrics to the server at each iteration using NVFlare's metric s
 ```bash
 python job.py --job_name "flwr-pt-tb" --content_dir "./flwr-pt-tb" --stream_metrics
 ```
+
+You can visualize the metrics streamed to the server using TensorBoard.
+```bash
+tensorboard --logdir /tmp/nvflare/hello-flower
+```
+![tensorboard training curve](./train.png)
+
+## Notes
+Make sure your `pyproject.toml` files in the Flower apps contain an "address" field. This needs to be present as the `--federation-config` option of the `flwr run` command tries to override the `“address”` field.
+Your `pyproject.toml` should include a section similar to this:
+```
+[tool.flwr.federations]
+default = "xxx"
+
+[tool.flwr.federations.xxx]
+options.num-supernodes = 2
+address = "127.0.0.1:9093"
+insecure = false
+```
+The number `options.num-supernodes` should match the number of NVFlare clients defined in [job.py](./job.py), e.g., `job.simulator_run(args.workdir, gpu="0", n_clients=2)`.

examples/hello-world/hello-flower/flwr-pt-tb/flwr_pt_tb/client.py

@@ -28,8 +28,6 @@
 # initializes NVFlare interface
 from nvflare.client.tracking import SummaryWriter
 
-flare.init()
-
 
 # Define FlowerClient and client_fn
 class FlowerClient(NumPyClient):
@@ -81,3 +79,15 @@ def client_fn(context: Context):
 app = ClientApp(
     client_fn=client_fn,
 )
+
+
+@app.enter()
+def enter(ctxt: Context) -> None:
+    flare.init()
+    print("ClientApp entering. Flare initialized.")
+
+
+@app.exit()
+def exit(ctxt: Context) -> None:
+    flare.shutdown()
+    print("ClientApp exiting. Flare shutdown.")

examples/hello-world/hello-flower/flwr-pt-tb/pyproject.toml

@@ -8,8 +8,8 @@ version = "1.0.0"
 description = ""
 license = "Apache-2.0"
 dependencies = [
-    "flwr[simulation]>=1.11.0,<2.0",
-    "nvflare~=2.5.0rc",
+    "flwr[simulation]>=1.15.2,<2.0",
+    "nvflare~=2.6.0rc",
     "torch==2.2.1",
     "torchvision==0.17.1",
     "tensorboard"
@@ -33,3 +33,5 @@ default = "local-simulation"
 
 [tool.flwr.federations.local-simulation]
 options.num-supernodes = 2
+address = "127.0.0.1:9093"
+insecure = true

examples/hello-world/hello-flower/flwr-pt/pyproject.toml

@@ -8,10 +8,11 @@ version = "1.0.0"
 description = ""
 license = "Apache-2.0"
 dependencies = [
-    "flwr[simulation]>=1.11.0,<2.0",
-    "nvflare~=2.5.0rc",
+    "flwr[simulation]>=1.15.2,<2.0",
+    "nvflare~=2.6.0rc",
     "torch==2.2.1",
     "torchvision==0.17.1",
+    "tensorboard"
 ]
 
 [tool.hatch.build.targets.wheel]
@@ -32,3 +33,5 @@ default = "local-simulation"
 
 [tool.flwr.federations.local-simulation]
 options.num-supernodes = 2
+address = "127.0.0.1:9093"
+insecure = true

examples/tutorials/self-paced-training/part-1_federated_learning_introduction/chapter-2_develop_federated_learning_applications/02.1_federated_statistics/federated_statistics_introduction.ipynb

@@ -0,0 +1,23 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Federated Statistics Introduction\n",
+    "\n",
+    "In a federated learning setting, because data is private at each site and we need to ensure data privacy, there are many considerations to take into account when trying to gather statistics on the data. We provide two examples, one for image data and one for tabular data:\n",
+    "\n",
+    "   * [Federated Statistics with image data](./federated_statistics_with_image_data/federated_statistics_with_image_data.ipynb) shows how to compute local and global image statistics with the consideration that data is private at each of the client sites.\n",
+    "   * [Federated Statistics with tabular data](./federated_statistics_with_tabular_data/federated_statistics_with_tabular_data.ipynb) demonstrates how to create federated statistics for data that can be represented as Pandas DataFrames."
+   ]
+  }
+ ],
+ "metadata": {
+  "language_info": {
+   "name": "python"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

examples/tutorials/self-paced-training/part-1_federated_learning_introduction/chapter-2_develop_federated_learning_applications/02.2_convert_torch_lightning_to_federated_learning/convert_torch_lightning_to_fl.ipynb

@@ -424,7 +424,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
+   "display_name": "Python 3",
    "language": "python",
    "name": "python3"
   },
@@ -438,7 +438,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.14"
+   "version": "3.10.12"
   }
  },
  "nbformat": 4,

examples/tutorials/self-paced-training/part-1_federated_learning_introduction/chapter-2_develop_federated_learning_applications/02.3_convert_machine_learning_to_federated_learning/convert_ml_to_fl.ipynb

@@ -0,0 +1,34 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Simple ML/DL to FL transition with NVFlare\n",
+    "\n",
+    "Converting Deep Learning (DL) models to Federated Learning (FL) entails several key steps:\n",
+    "\n",
+    " - Formulating the algorithm: This involves determining how to adapt a DL model into an FL framework, including specifying the information exchange protocol between the server and clients.\n",
+    "\n",
+    " - Code conversion: Adapting existing standalone DL code into FL-compatible code. This typically involves minimal changes, often just a few lines of code, thanks to tools like NVFlare.\n",
+    "\n",
+    " - Workflow configuration: Once the code is modified, configuring the workflow to integrate the newly adapted FL code seamlessly.\n",
+    "\n",
+    "NVFlare simplifies the process of transitioning from traditional Machine Learning (ML) or DL algorithms to FL. With NVFlare, the conversion process requires only minor code adjustments.\n",
+    "\n",
+    "In this section, we have the following three examples for converting traditional ML to FL:\n",
+    "\n",
+    "   * [Convert Logistics Regression to federated learning](02.3.1_convert_logistic_regression_to_federated_learning/convert_logistic_regression_to_fl.ipynb)\n",
+    "   * [Convert KMeans to federated learning](02.3.2_convert_kmeans_to_federated_learning/convert_kmeans_to_fl.ipynb)\n",
+    "   * [Convert Survival Analysis to federated learning](02.3.3_convert_survival_analysis_to_federated_learning/convert_survival_analysis_to_fl.ipynb)"
+   ]
+  }
+ ],
+ "metadata": {
+  "language_info": {
+   "name": "python"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}