Rename fabric run model to fabric run

docs/source-fabric/fundamentals/launch.rst

@@ -67,7 +67,7 @@ An alternative way to launch your Python script in multiple processes is to use
 
 .. code-block:: bash
 
-    fabric run model path/to/your/script.py
+    fabric run path/to/your/script.py
 
 This is essentially the same as running ``python path/to/your/script.py``, but it also lets you configure the following settings externally without changing your code:
 
@@ -80,9 +80,9 @@ This is essentially the same as running ``python path/to/your/script.py``, but i
 
 .. code-block:: bash
 
-    fabric run model --help
+    fabric run --help
 
-    Usage: fabric run model [OPTIONS] SCRIPT [SCRIPT_ARGS]...
+    Usage: fabric run [OPTIONS] SCRIPT [SCRIPT_ARGS]...
 
       Run a Lightning Fabric script.
 
@@ -128,7 +128,7 @@ Here is how you run DDP with 8 GPUs and `torch.bfloat16 <https://pytorch.org/doc
 
 .. code-block:: bash
 
-    fabric run model ./path/to/train.py \
+    fabric run ./path/to/train.py \
         --strategy=ddp \
         --devices=8 \
         --accelerator=cuda \
@@ -138,7 +138,7 @@ Or `DeepSpeed Zero3 <https://www.deepspeed.ai/2021/03/07/zero3-offload.html>`_ w
 
 .. code-block:: bash
 
-     fabric run model ./path/to/train.py \
+     fabric run ./path/to/train.py \
         --strategy=deepspeed_stage_3 \
         --devices=8 \
         --accelerator=cuda \
@@ -148,7 +148,7 @@ Or `DeepSpeed Zero3 <https://www.deepspeed.ai/2021/03/07/zero3-offload.html>`_ w
 
 .. code-block:: bash
 
-    fabric run model ./path/to/train.py \
+    fabric run ./path/to/train.py \
         --devices=auto \
         --accelerator=auto \
         --precision=16

docs/source-fabric/fundamentals/precision.rst

@@ -66,7 +66,7 @@ The same values can also be set through the :doc:`command line interface <launch
 
 .. code-block:: bash
 
-    lightning run model train.py --precision=bf16-mixed
+    fabric run train.py --precision=bf16-mixed
 
 
 .. note::

docs/source-fabric/guide/multi_node/barebones.rst

@@ -72,7 +72,7 @@ Log in to the **first node** and run this command:
 .. code-block:: bash
     :emphasize-lines: 2,3
 
-    lightning run model \
+    fabric run \
         --node-rank=0  \
         --main-address=10.10.10.16 \
         --accelerator=cuda \
@@ -85,7 +85,7 @@ Log in to the **second node** and run this command:
 .. code-block:: bash
     :emphasize-lines: 2,3
 
-    lightning run model \
+    fabric run \
         --node-rank=1  \
         --main-address=10.10.10.16 \
         --accelerator=cuda \
@@ -129,7 +129,7 @@ The most likely reasons and how to fix it:
 
     export GLOO_SOCKET_IFNAME=eno1
     export NCCL_SOCKET_IFNAME=eno1
-    lightning run model ...
+    fabric run ...
 
   You can find the interface name by parsing the output of the ``ifconfig`` command.
   The name of this interface **may differ on each node**.
@@ -152,7 +152,7 @@ Launch your command by prepending ``NCCL_DEBUG=INFO`` to get more info.
 
 .. code-block:: bash
 
-    NCCL_DEBUG=INFO lightning run model ...
+    NCCL_DEBUG=INFO fabric run ...
 
 
 ----

examples/fabric/image_classifier/README.md

@@ -27,11 +27,11 @@ This script shows you how to scale the pure PyTorch code to enable GPU and multi
 
 ```bash
 # CPU
-lightning run model train_fabric.py
+fabric run train_fabric.py
 
 # GPU (CUDA or M1 Mac)
-lightning run model train_fabric.py --accelerator=gpu
+fabric run train_fabric.py --accelerator=gpu
 
 # Multiple GPUs
-lightning run model train_fabric.py --accelerator=gpu --devices=4
+fabric run train_fabric.py --accelerator=gpu --devices=4
 ```

examples/fabric/image_classifier/train_fabric.py

@@ -20,10 +20,10 @@
 3. Apply ``setup`` over each model and optimizers pair, ``setup_dataloaders`` on all your dataloaders,
 and replace ``loss.backward()`` with ``self.backward(loss)``.
 
-4. Run the script from the terminal using ``lightning run model path/to/train.py``
+4. Run the script from the terminal using ``fabric run path/to/train.py``
 
 Accelerate your training loop by setting the ``--accelerator``, ``--strategy``, ``--devices`` options directly from
-the command line. See ``lightning run model --help`` or learn more from the documentation:
+the command line. See ``fabric run --help`` or learn more from the documentation:
 https://lightning.ai/docs/fabric.
 
 """
@@ -71,7 +71,7 @@ def forward(self, x):
 
 def run(hparams):
     # Create the Lightning Fabric object. The parameters like accelerator, strategy, devices etc. will be proided
-    # by the command line. See all options: `lightning run model --help`
+    # by the command line. See all options: `fabric run --help`
     fabric = Fabric()
 
     seed_everything(hparams.seed)  # instead of torch.manual_seed(...)
@@ -168,7 +168,7 @@ def run(hparams):
 if __name__ == "__main__":
     # Arguments can be passed in through the CLI as normal and will be parsed here
     # Example:
-    # lightning run model image_classifier.py accelerator=cuda --epochs=3
+    # fabric run image_classifier.py accelerator=cuda --epochs=3
     parser = argparse.ArgumentParser(description="Fabric MNIST Example")
     parser.add_argument(
         "--batch-size", type=int, default=64, metavar="N", help="input batch size for training (default: 64)"

examples/fabric/kfold_cv/README.md

@@ -14,13 +14,13 @@ This script shows you how to scale the pure PyTorch code to enable GPU and multi
 
 ```bash
 # CPU
-lightning run model train_fabric.py
+fabric run train_fabric.py
 
 # GPU (CUDA or M1 Mac)
-lightning run model train_fabric.py --accelerator=gpu
+fabric run train_fabric.py --accelerator=gpu
 
 # Multiple GPUs
-lightning run model train_fabric.py --accelerator=gpu --devices=4
+fabric run train_fabric.py --accelerator=gpu --devices=4
 ```
 
 ### References

examples/fabric/kfold_cv/train_fabric.py

@@ -107,7 +107,7 @@ def validate_dataloader(model, data_loader, fabric, hparams, fold, acc_metric):
 
 def run(hparams):
     # Create the Lightning Fabric object. The parameters like accelerator, strategy, devices etc. will be proided
-    # by the command line. See all options: `lightning run model --help`
+    # by the command line. See all options: `fabric run --help`
     fabric = Fabric()
 
     seed_everything(hparams.seed)  # instead of torch.manual_seed(...)
@@ -171,7 +171,7 @@ def run(hparams):
 if __name__ == "__main__":
     # Arguments can be passed in through the CLI as normal and will be parsed here
     # Example:
-    # lightning run model image_classifier.py accelerator=cuda --epochs=3
+    # fabric run image_classifier.py accelerator=cuda --epochs=3
     parser = argparse.ArgumentParser(description="Fabric MNIST K-Fold Cross Validation Example")
     parser.add_argument(
         "--batch-size", type=int, default=64, metavar="N", help="input batch size for training (default: 64)"

examples/fabric/language_model/README.md

@@ -7,11 +7,11 @@ It is a simplified version of the [official PyTorch example](https://github.com/
 
 ```bash
 # CPU
-lightning run model --accelerator=cpu train.py
+fabric run --accelerator=cpu train.py
 
 # GPU (CUDA or M1 Mac)
-lightning run model --accelerator=gpu train.py
+fabric run --accelerator=gpu train.py
 
 # Multiple GPUs
-lightning run model --accelerator=gpu --devices=4 train.py
+fabric run --accelerator=gpu --devices=4 train.py
 ```

examples/fabric/meta_learning/README.md

@@ -33,7 +33,7 @@ torchrun --nproc_per_node=2 --standalone train_torch.py
 **Accelerated using Lightning Fabric:**
 
 ```bash
-lightning run model train_fabric.py --devices 2 --strategy ddp --accelerator cpu
+fabric run train_fabric.py --devices 2 --strategy ddp --accelerator cpu
 ```
 
 ### References

examples/fabric/meta_learning/train_fabric.py

@@ -12,7 +12,7 @@
 - gym<=0.22
 
 Run it with:
-    lightning run model train_fabric.py --accelerator=cuda --devices=2 --strategy=ddp
+    fabric run train_fabric.py --accelerator=cuda --devices=2 --strategy=ddp
 """
 
 import cherry
@@ -59,7 +59,7 @@ def main(
     seed=42,
 ):
     # Create the Fabric object
-    # Arguments get parsed from the command line, see `lightning run model --help`
+    # Arguments get parsed from the command line, see `fabric run --help`
     fabric = Fabric()
 
     meta_batch_size = meta_batch_size // fabric.world_size

examples/fabric/reinforcement_learning/README.md

@@ -40,7 +40,7 @@ torchrun --nproc_per_node=2 --standalone train_torch.py
 ### Lightning Fabric:
 
 ```bash
-lightning run model --accelerator=cpu --strategy=ddp --devices=2 train_fabric.py
+fabric run --accelerator=cpu --strategy=ddp --devices=2 train_fabric.py
 ```
 
 ### Visualizing logs
@@ -71,7 +71,7 @@ The following video shows a trained agent on the [LunarLander-v2 environment](ht
 The agent was trained with the following:
 
 ```bash
-lightning run model \
+fabric run \
   --accelerator=cpu \
   --strategy=ddp \
   --devices=2 \
@@ -98,25 +98,25 @@ where, differently from the previous example, we have completely decoupled the e
 So for example:
 
 ```bash
-lightning run model --devices=3 train_fabric_decoupled.py --num-envs 4
+fabric run --devices=3 train_fabric_decoupled.py --num-envs 4
 ```
 
 will spawn 3 processes, one is the Player and the others the Trainers, with the Player running 4 independent environments, where every process runs on the CPU;
 
 ```bash
-lightning run model --devices=3 train_fabric_decoupled.py --num-envs 4 --cuda
+fabric run --devices=3 train_fabric_decoupled.py --num-envs 4 --cuda
 ```
 
 will instead run only the Trainers on the GPU.
 If one wants to run both the Player and the Trainers on the GPU, then both the flags `--cuda` and `--player-on-gpu` must be provided:
 
 ```bash
-lightning run model --devices=3 train_fabric_decoupled.py --num-envs 4 --cuda --player-on-gpu
+fabric run --devices=3 train_fabric_decoupled.py --num-envs 4 --cuda --player-on-gpu
 ```
 
 > **Warning**
 >
-> With this second example, there is no need for the user to provide the `accelerator` and the `strategy` to the `lightning run model` script.
+> With this second example, there is no need for the user to provide the `accelerator` and the `strategy` to the `fabric run` script.
 
 ## Number of updates, environment steps and share data
 

examples/fabric/reinforcement_learning/train_fabric.py

@@ -14,7 +14,7 @@
 
 
 Run it with:
-    lightning run model --accelerator=cpu --strategy=ddp --devices=2 train_fabric.py
+    fabric run --accelerator=cpu --strategy=ddp --devices=2 train_fabric.py
 """
 
 import argparse

examples/fabric/reinforcement_learning/train_fabric_decoupled.py

@@ -14,7 +14,7 @@
 
 
 Run it with:
-    lightning run model --devices=2 train_fabric_decoupled.py
+    fabric run --devices=2 train_fabric_decoupled.py
 """
 
 import argparse

src/lightning/app/cli/lightning_cli.py

@@ -18,7 +18,6 @@
 from typing import Tuple, Union
 
 import click
-from lightning_utilities.core.imports import RequirementCache
 from requests.exceptions import ConnectionError
 
 import lightning.app.core.constants as constants
@@ -303,13 +302,6 @@ def run_app(
     )
 
 
-if RequirementCache("lightning-fabric>=1.9.0") or RequirementCache("lightning>=1.9.0"):
-    # note it is automatically replaced to `from lightning.fabric.cli` when building monolithic/mirror package
-    from lightning.fabric.cli import _run_model
-
-    run.add_command(_run_model)
-
-
 @_main.command("open", hidden=True)
 @click.argument("path", type=str, default=".")
 @click.option("--name", help="The name to use for the CloudSpace", default="", type=str)

src/lightning/fabric/CHANGELOG.md

@@ -17,7 +17,7 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 ### Changed
 
-- Renamed `lightning run model` to `fabric run model` ([#19442](https://github.com/Lightning-AI/pytorch-lightning/pull/19442))
+- Renamed `lightning run model` to `fabric run` ([#19442](https://github.com/Lightning-AI/pytorch-lightning/pull/19442), [#19527](https://github.com/Lightning-AI/pytorch-lightning/pull/19527))
 
 
 - The `Fabric.rank_zero_first` context manager now uses a barrier without timeout to avoid long-running tasks to be interrupted ([#19448](https://github.com/Lightning-AI/lightning/pull/19448))

src/lightning/fabric/cli.py

@@ -55,7 +55,7 @@ def _legacy_main() -> None:
         """
         print(
             "`lightning run model` is deprecated and will be removed in future versions."
-            " Please call `fabric run model` instead."
+            " Please call `fabric run` instead."
         )
         args = sys.argv[1:]
         if args and args[0] == "run" and args[1] == "model":
@@ -70,12 +70,8 @@ def _legacy_main() -> None:
     def _main() -> None:
         pass
 
-    @_main.group()
-    def run() -> None:
-        pass
-
-    @run.command(
-        "model",
+    @_main.command(
+        "run",
         context_settings={
             "ignore_unknown_options": True,
         },
@@ -146,7 +142,7 @@ def run() -> None:
         ),
     )
     @click.argument("script_args", nargs=-1, type=click.UNPROCESSED)
-    def _run_model(**kwargs: Any) -> None:
+    def _run(**kwargs: Any) -> None:
         """Run a Lightning Fabric script.
 
         SCRIPT is the path to the Python script with the code to run. The script must contain a Fabric object.
@@ -225,4 +221,4 @@ def main(args: Namespace, script_args: Optional[List[str]] = None) -> None:
         )
         raise SystemExit(1)
 
-    _run_model()
+    _run()

src/lightning/fabric/fabric.py

@@ -839,7 +839,7 @@ def launch(self, function: Callable[["Fabric"], Any] = _do_nothing, *args: Any,
             Returns the output of the function that ran in worker process with rank 0.
 
         The ``launch()`` method should only be used if you intend to specify accelerator, devices, and so on in
-        the code (programmatically). If you are launching with the Lightning CLI, ``lightning run model ...``, remove
+        the code (programmatically). If you are launching with the Lightning CLI, ``fabric run ...``, remove
         ``launch()`` from your code.
 
         The ``launch()`` is a no-op when called multiple times and no function is passed in.
@@ -1028,7 +1028,7 @@ def _validate_launched(self) -> None:
         if not self._launched and not isinstance(self._strategy, (SingleDeviceStrategy, DataParallelStrategy)):
             raise RuntimeError(
                 "To use Fabric with more than one device, you must call `.launch()` or use the CLI:"
-                " `lightning run model --help`."
+                " `fabric run --help`."
             )
 
     def _validate_setup(self, module: nn.Module, optimizers: Sequence[Optimizer]) -> None: