Remove convert_to_half, suggest using model.half (#7974)

Author: Sean Naren

pytorch_lightning/plugins/training_type/ipu.py

@@ -19,7 +19,6 @@
 import torch
 from torch.utils.data import DataLoader
 
-from pytorch_lightning import _logger as log
 from pytorch_lightning.callbacks import GradientAccumulationScheduler
 from pytorch_lightning.core.lightning import LightningModule
 from pytorch_lightning.overrides.base import _LightningModuleWrapperBase
@@ -67,7 +66,6 @@ def __init__(
         device_iterations: int = 1,
         autoreport: bool = True,
         autoreport_dir: Optional[str] = None,
-        convert_model_to_half: bool = False,
         parallel_devices: Optional[List[torch.device]] = None,
         cluster_environment: Optional[ClusterEnvironment] = None,
         training_opts: Optional['poptorch.Options'] = None,
@@ -82,7 +80,6 @@ def __init__(
             autoreport: Enable auto-reporting for IPUs using PopVision
                 https://docs.graphcore.ai/projects/graphcore-popvision-user-guide/en/latest/graph/graph.html
             autoreport_dir: Optional directory to store autoReport output.
-            convert_model_to_half: Converts the model to half precision, which can be used for pure FP16 training.
             training_opts: Optional ``poptorch.Options`` to override the default created options for training.
             inference_opts: Optional ``poptorch.Options`` to override the default
                 created options for validation/testing and predicting.
@@ -94,7 +91,6 @@ def __init__(
                 "Learn more or get started with IPUs at https://www.graphcore.ai/getstarted"
             )
 
-        self.convert_model_to_half = convert_model_to_half
         self.device_iterations = device_iterations
         self.autoreport = autoreport
         self.autoreport_dir = autoreport_dir
@@ -113,12 +109,7 @@ def __init__(
 
     def pre_dispatch(self) -> None:
         self._handle_gradient_accumulation_steps()
-        if self.convert_model_to_half:
-            log.info('Using full 16bit precision, converting LightningModule weights to FP16.')
-            self.model = self.model.half()
         precision = self.lightning_module.trainer.precision
-        precision = 16 if self.convert_model_to_half else precision
-
         model = LightningIPUModule(self.lightning_module, precision)
         self.model = model
 

tests/accelerators/test_ipu.py

@@ -200,20 +200,13 @@ class TestCallback(Callback):
 
         def on_train_start(self, trainer: Trainer, pl_module: LightningModule) -> None:
             assert trainer.accelerator.model.precision == 16
-            assert trainer.accelerator.training_type_plugin.convert_model_to_half
             for param in trainer.accelerator.model.parameters():
                 assert param.dtype == torch.float16
             raise SystemExit
 
     model = IPUModel()
-    trainer = Trainer(
-        default_root_dir=tmpdir,
-        fast_dev_run=True,
-        ipus=1,
-        precision=16,
-        plugins=IPUPlugin(convert_model_to_half=True),
-        callbacks=TestCallback()
-    )
+    model = model.half()
+    trainer = Trainer(default_root_dir=tmpdir, fast_dev_run=True, ipus=1, precision=16, callbacks=TestCallback())
 
     assert isinstance(trainer.accelerator.training_type_plugin, IPUPlugin)
     assert isinstance(trainer.accelerator.precision_plugin, IPUPrecisionPlugin)