Update Gradient Clipping for TPU Accelerator (#6576)

CHANGELOG.md

@@ -167,6 +167,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Fixed a bug where `all_gather` would not work correctly with `tpu_cores=8` ([#6587](https://github.com/PyTorchLightning/pytorch-lightning/pull/6587))
 
 
+- Update Gradient Clipping for the TPU Accelerator ([#6576](https://github.com/PyTorchLightning/pytorch-lightning/pull/6576))
+
+
 ## [1.2.3] - 2021-03-09
 
 ### Fixed

pytorch_lightning/accelerators/tpu.py

@@ -1,4 +1,4 @@
-from typing import Any, Callable, Optional, TYPE_CHECKING
+from typing import Any, Callable, Optional, TYPE_CHECKING, Union
 
 import torch
 from torch.optim import Optimizer
@@ -12,6 +12,9 @@ from pytorch_lightning.utilities.exceptions import MisconfigurationException
 
 if _XLA_AVAILABLE:
     import torch_xla.core.xla_model as xm
+    from torch_xla._patched_functions import clip_grad_norm_
+
+    xla_clip_grad_norm_ = clip_grad_norm_
 
 if TYPE_CHECKING:
     from pytorch_lightning.core.lightning import LightningModule
@@ -55,3 +58,16 @@ class TPUAccelerator(Accelerator):
         if isinstance(self.training_type_plugin, TPUSpawnPlugin) and self.training_type_plugin.is_distributed:
             return xm.all_gather(tensor).view(-1, *tensor.shape)
         return tensor
+
+    def clip_gradients(self, optimizer: Optimizer, clip_val: Union[float, int], norm_type: float = 2.0):
+
+        model = self.lightning_module
+        parameters = model.parameters()
+
+        grad_clip_val = float(clip_val)
+        if grad_clip_val <= 0:
+            return
+
+        max_norm = grad_clip_val
+
+        xla_clip_grad_norm_(parameters, max_norm, norm_type)

pytorch_lightning/plugins/precision/precision_plugin.py

@@ -100,7 +100,6 @@ class PrecisionPlugin(Plugin):
 
     def clip_gradients(self, optimizer: 'Optimizer', clip_val: Union[int, float], norm_type: float = 2.0) -> None:
         """Clips the gradients to a specific value"""
-        # TODO: separate TPU case from here
         if clip_val is None:
             return
 

tests/models/test_tpu.py

@@ -355,3 +355,31 @@ def test_tpu_reduce():
                 assert result.item() == 8
 
     xmp.spawn(test_reduce, nprocs=8, start_method='fork')
+
+
+@pytest.mark.parametrize("clip_val", [0, 10])
+@RunIf(tpu=True)
+@pl_multi_process_test
+@mock.patch("pytorch_lightning.accelerators.tpu.xla_clip_grad_norm_")
+def test_tpu_precision_16_clip_gradients(mock_clip_grad_norm, clip_val, tmpdir):
+    """
+    Ensure that clip gradients is only called if the value is greater than 0.
+    """
+    tutils.reset_seed()
+    trainer_options = dict(
+        default_root_dir=tmpdir,
+        progress_bar_refresh_rate=0,
+        max_epochs=1,
+        tpu_cores=1,
+        precision=16,
+        limit_train_batches=4,
+        limit_val_batches=4,
+        gradient_clip_val=clip_val,
+    )
+    model = BoringModel()
+    tpipes.run_model_test(trainer_options, model, on_gpu=False, with_hpc=False)
+
+    if clip_val > 0:
+        mock_clip_grad_norm.assert_called()
+    else:
+        mock_clip_grad_norm.assert_not_called()