Modified and Added tutorial Workflow-Interface_201_Exclusive_GPUs_wit…

…h_Ray.ipynb Signed-off-by: Parth Mandaliya <[email protected]>
securefederatedai · May 5, 2023 · 8b17d7a · 8b17d7a
1 parent eadfd1e
commit 8b17d7a
Show file tree

Hide file tree

Showing 7 changed files with 132 additions and 57 deletions.
diff --git a/openfl-tutorials/experimental/Privacy_Meter/cifar10_PM.py b/openfl-tutorials/experimental/Privacy_Meter/cifar10_PM.py
@@ -712,6 +712,8 @@ def end(self):
     )
 
     # Split train, test, and population dataset among collaborators
+    # this function will be called before executing collaborator steps
+    # which will return private attributes dictionary for each collaborator
     def callable_to_initialize_collaborator_private_attributes(
             index, n_collaborators, train_ds, test_ds, population_ds, args):
         # construct the training and test and population dataset
@@ -778,17 +780,17 @@ def callable_to_initialize_collaborator_private_attributes(
             Collaborator(
                 name=collab_name,
                 private_attributes_callable=callable_to_initialize_collaborator_private_attributes,
-                # Set `num_gpus=0.5` to `num_gpus=0.0` to run on CPU
-                num_cpus=0.0, num_gpus=0.5,  # Assuming GPU is available in the machine
+                # If 1 GPU is available in the machine
+                # Set `num_gpus=0.0` to `num_gpus=0.5` to run on GPU with ray backend with 2 collaborators
+                num_cpus=0.0, num_gpus=0.0,
                 index=idx, n_collaborators=len(collaborator_names),
                 train_ds=train_dataset, test_ds=test_dataset,
                 population_ds=population_dataset, args=args
             )
         )
 
-    # To activate the ray backend with parallel collaborator tasks run in their own process
-    # and exclusive GPUs assigned to tasks, set LocalRuntime with backend='ray':
-    local_runtime = LocalRuntime(aggregator=aggregator, collaborators=collaborators)
+    # Set backend='ray' to use ray-backend
+    local_runtime = LocalRuntime(aggregator=aggregator, collaborators=collaborators, backend="single_process")
 
     print(f"Local runtime collaborators = {local_runtime.collaborators}")
 

diff --git a/openfl-tutorials/experimental/Workflow_Interface_101_MNIST.ipynb b/openfl-tutorials/experimental/Workflow_Interface_101_MNIST.ipynb
@@ -1,6 +1,7 @@
 {
  "cells": [
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "14821d97",
    "metadata": {},
@@ -10,6 +11,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "bd059520",
    "metadata": {},
@@ -23,6 +25,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "39c3d86a",
    "metadata": {},
@@ -31,6 +34,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "a7989e72",
    "metadata": {},
@@ -39,6 +43,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "fc8e35da",
    "metadata": {},
@@ -47,6 +52,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "4dbb89b6",
    "metadata": {},
@@ -72,6 +78,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "7237eac4",
    "metadata": {},
@@ -165,6 +172,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "cd268911",
    "metadata": {},
@@ -312,9 +320,9 @@
    "id": "2aabf61e",
    "metadata": {},
    "source": [
-    "You'll notice in the `FederatedFlow` definition above that there were certain attributes that the flow was not initialized with, namely the `train_loader` and `test_loader` for each of the collaborators. These are **private_attributes** that are exposed only through the runtime. Each participant has it's own set of private attributes: a dictionary where the key is the attribute name, and the value is the object that will be made accessible through that participant's task. \n",
+    "Note that the private attributes are flexible, and you can choose to pass in a completely different type of object to any of the collaborators or aggregator (with an arbitrary name). These private attributes will always be filtered out of the current state when transferring from collaborator to aggregator, or vice versa.  \n",
     "\n",
-    "Below, we segment shards of the MNIST dataset for **four collaborators**: Portland, Seattle, Chandler, and Portland. Each has their own slice of the dataset that is accessible through the `train_loader` and `test_loader` attributes, which are set using the `callable_to_initialize_collaborator_private_attributes` callable function. Note that the private attributes are flexible, and you can choose to pass in a completely different type of object to any of the collaborators or aggregator (with an arbitrary name). These private attributes will always be filtered out of the current state when transfering from collaborator to aggregator, or vice versa.  "
+    "Private attributes can be set using callback function while instantiating the participant. Parameters required by the callback function are specified as arguments while instantiating the participant. In this example callback function, `callable_to_initialize_collaborator_private_attributes`, returns the private attributes `train_loader` and `test_loader` of the collaborator. Parameters required by the callback function `index`, `n_collaborators`, `batch_size`, `train_dataset`, `test_dataset` are passed appropriate values with the same names in the Collaborator constructor."
    ]
   },
   {
@@ -356,8 +364,25 @@
     "\n",
     "local_runtime = LocalRuntime(aggregator=aggregator_, collaborators=collaborators,\n",
     "                             backend=\"ray\")\n",
-    "print(f'Local runtime collaborators = {local_runtime.collaborators}')\n",
-    "\n",
+    "print(f'Local runtime collaborators = {local_runtime.collaborators}')"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "id": "278ad46b",
+   "metadata": {},
+   "source": [
+    "Now that we have our flow and runtime defined, let's run the experiment! "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a175b4d6",
+   "metadata": {},
+   "outputs": [],
+   "source": [
     "model = None\n",
     "best_model = None\n",
     "optimizer = None\n",
@@ -367,11 +392,12 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
-   "id": "278ad46b",
+   "id": "86b3dd2e",
    "metadata": {},
    "source": [
-    "Now that we have our flow and runtime defined, let's run the experiment! "
+    "Now that the flow has completed, let's get the final model and accuracy"
    ]
   },
   {
@@ -387,6 +413,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "5dd1558c",
    "metadata": {},
@@ -409,6 +436,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "a61a876d",
    "metadata": {},
@@ -458,6 +486,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "b55ccb19",
    "metadata": {},
@@ -486,6 +515,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "e5efa1ff",
    "metadata": {},
@@ -504,6 +534,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "3292b2e0",
    "metadata": {},
@@ -542,6 +573,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "eb1866b7",
    "metadata": {},
@@ -570,6 +602,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "ef877a50",
    "metadata": {},
@@ -598,6 +631,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "9826c45f",
    "metadata": {},
@@ -616,6 +650,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "dd962ddc",
    "metadata": {},
@@ -634,6 +669,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "426f2395",
    "metadata": {},

diff --git a/openfl-tutorials/experimental/Workflow_Interface_102_Aggregator_Validation.ipynb b/openfl-tutorials/experimental/Workflow_Interface_102_Aggregator_Validation.ipynb
@@ -1,6 +1,7 @@
 {
  "cells": [
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "14821d97",
    "metadata": {},
@@ -10,6 +11,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "bd059520",
    "metadata": {},
@@ -18,6 +20,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "fc8e35da",
    "metadata": {},
@@ -26,6 +29,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "4dbb89b6",
    "metadata": {},
@@ -50,6 +54,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "7237eac4",
    "metadata": {},
@@ -133,6 +138,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "cd268911",
    "metadata": {},
@@ -169,6 +175,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "b2e45614",
    "metadata": {
@@ -284,7 +291,9 @@
    "source": [
     "You'll notice in the `FederatedFlow` definition above that there were certain attributes that the flow was not initialized with, namely the `train_loader` and `test_loader` for each of the collaborators. These are **private_attributes** that are exposed only throught he runtime. Each participant has it's own set of private attributes: a dictionary where the key is the attribute name, and the value is the object that will be made accessible through that participant's task. \n",
     "\n",
-    "Below, we segment shards of the MNIST dataset for **four collaborators**: Portland, Seattle, Chandler, and Portland. Each has their own slice of the dataset that is accessible through the `train_loader` and `test_loader` attributes, which are set using the `callable_to_initialize_collaborator_private_attributes` callable function. Note that the private attributes are flexible, and you can choose to pass in a completely different type of object to any of the collaborators or aggregator (with an arbitrary name). These private attributes will always be filtered out of the current state when transfering from collaborator to aggregator, or vice versa.\n"
+    "Below, we segment shards of the MNIST dataset for **four collaborators**: `Portland`, `Seattle`, `Chandler`, and `Portland`. Each has their own slice of the dataset that is accessible through the `train_loader` and `test_loader` attributes, which are set using the `callable_to_initialize_collaborator_private_attributes` callable function. Note that the private attributes are flexible, and you can choose to pass in a completely different type of object to any of the collaborators or aggregator (with an arbitrary name). These private attributes will always be filtered out of the current state when transfering from collaborator to aggregator, or vice versa.\n",
+    "\n",
+    "Private attributes can be set using callback function while instantiating the participant. Parameters required by the callback function are specified as arguments while instantiating the participant. In this example callback function, `callable_to_initialize_collaborator_private_attributes`, returns the private attributes `train_loader` and `test_loader` of the collaborator. Callback function, `callable_to_initialize_aggregator_private_attributes`, returns the private attribute `test_loader` of the Aggregator."
    ]
   },
   {
@@ -342,6 +351,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "0525eaa9",
    "metadata": {},
@@ -365,6 +375,7 @@
    ]
   },
   {
+   "attachments": {},
    "cell_type": "markdown",
    "id": "8b9f8d25",
    "metadata": {},