add cyclic workflow

examples/hello-world/python_jobs/pt/client_api_pt_cyclic_cc.py

@@ -0,0 +1,54 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.  All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from code.net import Net
+
+from nvflare import FedJob
+from nvflare.app_common.workflows.fedavg import FedAvg
+from nvflare import ScriptExecutor
+from nvflare.app_common.ccwf import CyclicServerController, CyclicClientController
+from nvflare.app_opt.pt.file_model_persistor import PTFileModelPersistor
+from nvflare.app_common.ccwf.comps.simple_model_shareable_generator import SimpleModelShareableGenerator
+
+
+if __name__ == "__main__":
+    n_clients = 2
+    num_rounds = 3
+    train_script = "code/cifar10_fl.py"
+
+    job = FedJob(name="cifar10_fedavg", init_model=Net(), external_scripts=[train_script])  # TODO: use load/save model in FedAvg
+
+    controller = CyclicServerController(
+        num_rounds=num_rounds,
+        max_status_report_interval=300
+    )
+    job.to(controller, "server")
+
+    for i in range(n_clients):
+        executor = ScriptExecutor(
+            task_script_path=train_script,
+            task_script_args=""  # f"--batch_size 32 --data_path /tmp/data/site-{i}"
+        )
+        job.to(executor, f"site-{i}", gpu=0)
+
+        # Add client-side controller for cyclic workflow
+        executor = CyclicClientController()
+        job.to(executor, f"site-{i}", tasks=["cyclic_*"])
+
+        # In swarm learning, each client uses a model persistor and shareable_generator
+        job.to(PTFileModelPersistor(model=Net()), f"site-{i}", id="persistor")
+        job.to(SimpleModelShareableGenerator(), f"site-{i}", id="shareable_generator")
+
+    job.export_job("/tmp/nvflare/jobs/job_config")
+    job.simulator_run("/tmp/nvflare/jobs/workdir")

examples/hello-world/python_jobs/pt/client_api_pt_swarm.py

@@ -21,8 +21,6 @@
 from nvflare.app_opt.pt.file_model_persistor import PTFileModelPersistor
 from nvflare.app_common.aggregators.intime_accumulate_model_aggregator import InTimeAccumulateWeightedAggregator
 from nvflare.app_common.ccwf.comps.simple_model_shareable_generator import SimpleModelShareableGenerator
-from nvflare.app_common.launchers.subprocess_launcher import SubprocessLauncher
-from nvflare.app_opt.pt.client_api_launcher_executor import PTClientAPILauncherExecutor
 from nvflare.fuel.utils.constants import Mode
 from nvflare.fuel.utils.pipe.file_pipe import FilePipe
 
@@ -49,18 +47,6 @@
         )
         job.to(executor, f"site-{i}", gpu=0, tasks=["train", "validate", "submit_model"])
 
-        # executor = PTClientAPILauncherExecutor(
-        #     launcher_id="launcher",
-        #     pipe_id="pipe"
-        # )
-        # job.to(executor, f"site-{i}", gpu=0, tasks=["train", "validate", "submit_model"])
-        #
-        # launcher = SubprocessLauncher(
-        #     script=train_script,
-        #     launch_once=True
-        # )
-        # job.to(launcher, f"site-{i}", id="launcher")
-
         pipe = FilePipe(
             mode=Mode.PASSIVE,
             root_path="{WORKSPACE}/{JOB_ID}/{SITE_NAME}",

nvflare/fed_job.py

@@ -78,7 +78,7 @@ def to(self, obj: Any, target: str, tasks: List[str] = None, gpu: int = None, fi
             if target not in self._deploy_map:
                 self._deploy_map[target] = ExecutorApp(external_scripts=self._external_scripts)
                 self.clients.append(target)
-                if gpu is not None:
+                if gpu is not None:  # TODO: make sure GPUs are not added several times per client. Use dict as well?
                     self.gpus.append(str(gpu))
             self._deploy_map[target].add_executor(obj, tasks=tasks)
         else: