rename dgx cloud executor to runai executor

Signed-off-by: Hemil Desai <hemild@nvidia.com>

Author: hemildesai

docs/source/guides/execution.md

@@ -41,7 +41,7 @@ The packager support matrix is described below:
 | DockerExecutor | run.Packager, run.GitArchivePackager, run.PatternPackager, run.HybridPackager |
 | SlurmExecutor | run.Packager, run.GitArchivePackager, run.PatternPackager, run.HybridPackager |
 | SkypilotExecutor | run.Packager, run.GitArchivePackager, run.PatternPackager, run.HybridPackager |
-| DGXCloudExecutor | run.Packager, run.GitArchivePackager, run.PatternPackager, run.HybridPackager |
+| RunAIExecutor | run.Packager, run.GitArchivePackager, run.PatternPackager, run.HybridPackager |
 | LeptonExecutor   | run.Packager, run.GitArchivePackager, run.PatternPackager, run.HybridPackager |
 
 `run.Packager` is a passthrough base packager.
@@ -224,19 +224,17 @@ executor = your_skypilot_cluster(nodes=8, devices=8, container_image="your-nemo-
 
 As demonstrated in the examples, defining executors in Python offers great flexibility. You can easily mix and match things like common environment variables, and the separation of tasks from executors enables you to run the same configured task on any supported executor.
 
-#### DGXCloudExecutor
+#### RunAIExecutor
 
-The `DGXCloudExecutor` integrates with a DGX Cloud cluster's Run:ai API to launch distributed jobs. It uses REST API calls to authenticate, identify the target project and cluster, and submit the job specification.
-
-> **_WARNING:_** Currently, the `DGXCloudExecutor` is only supported when launching experiments *from* a pod running on the DGX Cloud cluster itself. Furthermore, this launching pod must have access to a Persistent Volume Claim (PVC) where the experiment/job directories will be created, and this same PVC must also be configured to be mounted by the job being launched.
+The `RunAIExecutor` integrates with the Run:ai API to launch distributed jobs. It uses REST API calls to authenticate, identify the target project and cluster, and submit the job specification.
 
 Here's an example configuration:
 
 ```python
-def your_dgx_executor(nodes: int, gpus_per_node: int, container_image: str):
-    # Ensure these are set correctly for your DGX Cloud environment
+def your_runai_executor(nodes: int, gpus_per_node: int, container_image: str):
+    # Ensure these are set correctly for your RunAI environment
     # You might fetch these from environment variables or a config file
-    base_url = "YOUR_DGX_CLOUD_API_ENDPOINT" # e.g., https://<cluster-name>.<domain>/api/v1
+    base_url = "YOUR_RUNAI_API_ENDPOINT" # e.g., https://<cluster-name>.<domain>/api/v1
     app_id = "YOUR_RUNAI_APP_ID"
     app_secret = "YOUR_RUNAI_APP_SECRET"
     project_name = "YOUR_RUNAI_PROJECT_NAME"
@@ -245,7 +243,7 @@ def your_dgx_executor(nodes: int, gpus_per_node: int, container_image: str):
     pvc_name = "your-pvc-k8s-name" # The Kubernetes name of the PVC
     pvc_mount_path = "/your_custom_path" # The path where the PVC will be mounted inside the container
 
-    executor = run.DGXCloudExecutor(
+    executor = run.RunAIExecutor(
         base_url=base_url,
         app_id=app_id,
         app_secret=app_secret,
@@ -261,11 +259,11 @@ def your_dgx_executor(nodes: int, gpus_per_node: int, container_image: str):
     return executor
 
 # Example usage:
-# executor = your_dgx_executor(nodes=4, gpus_per_node=8, container_image="your-nemo-image")
+# executor = your_runai_executor(nodes=4, gpus_per_node=8, container_image="your-nemo-image")
 
 ```
 
-For a complete end-to-end example using DGX Cloud with NeMo, refer to the [NVIDIA DGX Cloud NeMo End-to-End Workflow Example](https://docs.nvidia.com/dgx-cloud/run-ai/latest/nemo-e2e-example.html).
+For a complete end-to-end example using RunAI with NeMo, refer to the [NVIDIA RunAI NeMo End-to-End Workflow Example](https://docs.nvidia.com/dgx-cloud/run-ai/latest/nemo-e2e-example.html).
 
 #### LeptonExecutor
 

nemo_run/__init__.py

@@ -22,11 +22,11 @@
 from nemo_run.cli.lazy import LazyEntrypoint, lazy_imports
 from nemo_run.config import Config, ConfigurableMixin, Partial, Script
 from nemo_run.core.execution.base import Executor, ExecutorMacros, import_executor
-from nemo_run.core.execution.dgxcloud import DGXCloudExecutor
 from nemo_run.core.execution.docker import DockerExecutor
 from nemo_run.core.execution.launcher import FaultTolerance, SlurmRay, SlurmTemplate, Torchrun
 from nemo_run.core.execution.lepton import LeptonExecutor
 from nemo_run.core.execution.local import LocalExecutor
+from nemo_run.core.execution.runai import RunAIExecutor
 from nemo_run.core.execution.skypilot import SkypilotExecutor
 from nemo_run.core.execution.slurm import SlurmExecutor
 from nemo_run.core.packaging import GitArchivePackager, HybridPackager, Packager, PatternPackager
@@ -48,7 +48,7 @@
     "ConfigurableMixin",
     "DevSpace",
     "DockerExecutor",
-    "DGXCloudExecutor",
+    "RunAIExecutor",
     "dryrun_fn",
     "Executor",
     "import_executor",

nemo_run/core/execution/dgxcloud.py nemo_run/core/execution/runai.pynemo_run/core/execution/dgxcloud.py renamed to nemo_run/core/execution/runai.py

@@ -21,7 +21,7 @@
 logger = logging.getLogger(__name__)
 
 
-class DGXCloudState(Enum):
+class RunAIState(Enum):
     CREATING = "Creating"
     INITIALIZING = "Initializing"
     RESUMING = "Resuming"
@@ -39,14 +39,14 @@ class DGXCloudState(Enum):
 
 
 @dataclass(kw_only=True)
-class DGXCloudExecutor(Executor):
+class RunAIExecutor(Executor):
     """
-    Dataclass to configure a DGX Executor.
+    Dataclass to configure a RunAI Executor.
 
-    This executor integrates with a DGX cloud endpoint for launching jobs
+    This executor integrates with a RunAI cloud endpoint for launching jobs
     via a REST API. It acquires an auth token, identifies the project/cluster,
     and launches jobs with a specified command. It can be adapted to meet user
-    authentication and job-submission requirements on DGX.
+    authentication and job-submission requirements on RunAI.
     """
 
     base_url: str
@@ -168,29 +168,29 @@ def move_data(self, token: str, project_id: str, cluster_id: str, sleep: float =
 
         resp_json = resp.json()
         workload_id = resp_json["workloadId"]
-        status = DGXCloudState(resp_json["actualPhase"])
+        status = RunAIState(resp_json["actualPhase"])
 
-        logger.info(f"Successfully created data movement workload {workload_id} on DGXCloud")
+        logger.info(f"Successfully created data movement workload {workload_id} on RunAI")
 
         while status in [
-            DGXCloudState.PENDING,
-            DGXCloudState.CREATING,
-            DGXCloudState.INITIALIZING,
-            DGXCloudState.RUNNING,
+            RunAIState.PENDING,
+            RunAIState.CREATING,
+            RunAIState.INITIALIZING,
+            RunAIState.RUNNING,
         ]:
             time.sleep(sleep)
             status = self.status(workload_id)
             logger.debug(
                 f"Polling data movement workload {workload_id}'s status. Current status is: {status}"
             )
 
-        if status is not DGXCloudState.COMPLETED:
+        if status is not RunAIState.COMPLETED:
             raise RuntimeError(f"Failed to move data to PVC. Workload status is {status}")
 
         resp = self.delete_workload(token, workload_id)
         if resp.status_code >= 200 and resp.status_code < 300:
             logger.info(
-                "Successfully deleted data movement workload %s on DGXCloud with response code %d",
+                "Successfully deleted data movement workload %s on RunAI with response code %d",
                 workload_id,
                 resp.status_code,
             )
@@ -286,7 +286,7 @@ def nproc_per_node(self) -> int:
             return self.nprocs_per_node
         return 1
 
-    def status(self, job_id: str) -> Optional[DGXCloudState]:
+    def status(self, job_id: str) -> Optional[RunAIState]:
         url = f"{self.base_url}/workloads/{job_id}"
         token = self.get_auth_token()
         if not token:
@@ -296,10 +296,10 @@ def status(self, job_id: str) -> Optional[DGXCloudState]:
         headers = self._default_headers(token=token)
         response = requests.get(url, headers=headers)
         if response.status_code != 200:
-            return DGXCloudState("Unknown")
+            return RunAIState("Unknown")
 
         r_json = response.json()
-        return DGXCloudState(r_json["phase"])
+        return RunAIState(r_json["phase"])
 
     def cancel(self, job_id: str):
         # Retrieve the authentication token for the REST calls
@@ -315,7 +315,7 @@ def cancel(self, job_id: str):
         response = requests.get(url, headers=headers)
         if response.status_code >= 200 and response.status_code < 300:
             logger.info(
-                "Successfully cancelled job %s on DGX with response code %d",
+                "Successfully cancelled job %s on RunAI with response code %d",
                 job_id,
                 response.status_code,
             )
@@ -328,9 +328,9 @@ def cancel(self, job_id: str):
             )
 
     @classmethod
-    def logs(cls: Type["DGXCloudExecutor"], app_id: str, fallback_path: Optional[str]):
+    def logs(cls: Type["RunAIExecutor"], app_id: str, fallback_path: Optional[str]):
         logger.warning(
-            "Logs not available for DGXCloudExecutor based jobs. Please visit the cluster UI to view the logs."
+            "Logs not available for RunAIExecutor based jobs. Please visit the cluster UI to view the logs."
         )
 
     def cleanup(self, handle: str): ...

nemo_run/run/experiment.py

@@ -35,9 +35,8 @@
 from rich.console import Group
 from rich.live import Live
 from rich.panel import Panel
-from rich.progress import BarColumn, Progress, SpinnerColumn
+from rich.progress import BarColumn, Progress, SpinnerColumn, TaskID, TimeElapsedColumn
 from rich.progress import Task as RichTask
-from rich.progress import TaskID, TimeElapsedColumn
 from rich.syntax import Syntax
 from torchx.specs.api import AppState
 
@@ -51,10 +50,10 @@
     get_type_namespace,
 )
 from nemo_run.core.execution.base import Executor
-from nemo_run.core.execution.dgxcloud import DGXCloudExecutor
 from nemo_run.core.execution.docker import DockerExecutor
 from nemo_run.core.execution.lepton import LeptonExecutor
 from nemo_run.core.execution.local import LocalExecutor
+from nemo_run.core.execution.runai import RunAIExecutor
 from nemo_run.core.execution.skypilot import SkypilotExecutor
 from nemo_run.core.execution.slurm import SlurmExecutor
 from nemo_run.core.frontend.console.api import CONSOLE, configure_logging, deconfigure_logging
@@ -203,13 +202,13 @@ class Experiment(ConfigurableMixin):
         LocalExecutor,
         SkypilotExecutor,
         DockerExecutor,
-        DGXCloudExecutor,
+        RunAIExecutor,
         LeptonExecutor,
     )
     _DETACH_SUPPORTED_EXECUTORS = (
         SlurmExecutor,
         SkypilotExecutor,
-        DGXCloudExecutor,
+        RunAIExecutor,
         LeptonExecutor,
     )
     _DEPENDENCY_SUPPORTED_EXECUTORS = (SlurmExecutor,)

nemo_run/run/torchx_backend/components/ft_launcher.py

@@ -49,7 +49,7 @@ def ft_launcher(
     rank_termination_signal: Optional[str] = None,
     log_level: Optional[str] = None,
     max_restarts: Optional[int] = None,
-    dgxc: bool = False,
+    runai: bool = False,
     use_env: bool = False,
 ) -> specs.AppDef:
     torchrun_component = torchrun.torchrun(
@@ -71,7 +71,7 @@ def ft_launcher(
         mounts=mounts,
         debug=debug,
         max_retries=max_retries,
-        dgxc=dgxc,
+        runai=runai,
         use_env=use_env,
     )
 

nemo_run/run/torchx_backend/components/torchrun.py

@@ -60,7 +60,7 @@ def torchrun(
     rdzv_id: Optional[int] = None,
     mounts: Optional[list[str]] = None,
     debug: bool = False,
-    dgxc: bool = False,
+    runai: bool = False,
     lepton: bool = False,
     use_env: bool = False,
 ) -> specs.AppDef:
@@ -96,7 +96,7 @@ def torchrun(
         mounts: mounts to mount into the worker environment/container (ex. type=<bind/volume>,src=/host,dst=/job[,readonly]).
                 See scheduler documentation for more info.
         debug: whether to run with preset debug flags enabled
-        dgxc: whether to use a subset of settings for DGX Cloud
+        runai: whether to use a subset of settings for RunAI
         lepton: whether the experiment is running on Lepton AI
     """
     if (script is None) == (m is None):
@@ -140,7 +140,7 @@ def torchrun(
     if debug:
         env.update(_TORCH_DEBUG_FLAGS)
 
-    if dgxc:
+    if runai:
         cmd = ["--nnodes", nnodes_rep, "--nproc-per-node", nproc_per_node]
     elif lepton:
         cmd = [

nemo_run/run/torchx_backend/packaging.py

@@ -23,10 +23,10 @@
 
 from nemo_run.config import SCRIPTS_DIR, USE_WITH_RAY_CLUSTER_KEY, Partial, Script
 from nemo_run.core.execution.base import Executor
-from nemo_run.core.execution.dgxcloud import DGXCloudExecutor
 from nemo_run.core.execution.launcher import FaultTolerance, Torchrun
 from nemo_run.core.execution.lepton import LeptonExecutor
 from nemo_run.core.execution.local import LocalExecutor
+from nemo_run.core.execution.runai import RunAIExecutor
 from nemo_run.core.execution.slurm import SlurmExecutor
 from nemo_run.core.serialization.yaml import YamlSerializer
 from nemo_run.core.serialization.zlib_json import ZlibJSONSerializer
@@ -170,7 +170,7 @@ def _get_details_from_script(fn_or_script: Script, serialize_configs: bool):
             mounts=mounts,
             debug=executor.packager.debug,
             max_retries=executor.retries,
-            dgxc=isinstance(executor, DGXCloudExecutor),
+            runai=isinstance(executor, RunAIExecutor),
             lepton=isinstance(executor, LeptonExecutor),
             use_env=use_env,
         )

nemo_run/run/torchx_backend/schedulers/api.py

@@ -18,10 +18,10 @@
 from torchx.specs import AppDef, AppDryRunInfo
 
 from nemo_run.core.execution.base import Executor
-from nemo_run.core.execution.dgxcloud import DGXCloudExecutor
 from nemo_run.core.execution.docker import DockerExecutor
 from nemo_run.core.execution.lepton import LeptonExecutor
 from nemo_run.core.execution.local import LocalExecutor
+from nemo_run.core.execution.runai import RunAIExecutor
 from nemo_run.core.execution.skypilot import SkypilotExecutor
 from nemo_run.core.execution.slurm import SlurmExecutor
 
@@ -30,7 +30,7 @@
     SkypilotExecutor: "skypilot",
     LocalExecutor: "local_persistent",
     DockerExecutor: "docker_persistent",
-    DGXCloudExecutor: "dgx_cloud",
+    RunAIExecutor: "runai",
     LeptonExecutor: "lepton",
 }
 
@@ -39,7 +39,7 @@
     "skypilot": SkypilotExecutor,
     "local_persistent": LocalExecutor,
     "docker_persistent": DockerExecutor,
-    "dgx_cloud": DGXCloudExecutor,
+    "runai": RunAIExecutor,
     "lepton": LeptonExecutor,
 }