Apply suggestions from code review

Reverting changes to docstrings

Signed-off-by: angelinalg <[email protected]>

python/ray/train/examples/accelerate/accelerate_torch_trainer.py

@@ -25,7 +25,7 @@
 
 
 def train_func(config):
- """Your training function that is launched on each worker."""
+ """Your training function that will be launched on each worker."""
 
  # Unpack training configs
  lr = config["lr"]
@@ -116,7 +116,7 @@ def collate_fn(batch):
  eval_metric = metric.compute()
  accelerator.print(f"epoch {epoch}:", eval_metric)
 
- # Report checkpoint and metrics to Ray Train
+ # Report Checkpoint and metrics to Ray Train
  # ==========================================
  with TemporaryDirectory() as tmpdir:
  if accelerator.is_main_process:

python/ray/train/examples/pytorch/torch_fashion_mnist_example.py

@@ -19,23 +19,23 @@ def get_dataloaders(batch_size):
  transform = transforms.Compose([ToTensor(), Normalize((0.5,), (0.5,))])
 
  with FileLock(os.path.expanduser("~/data.lock")):
- # Download training data from open datasets
+ # Download training data from open datasets.
  training_data = datasets.FashionMNIST(
  root="~/data",
  train=True,
  download=True,
  transform=transform,
  )
 
- # Download test data from open datasets
+ # Download test data from open datasets.
  test_data = datasets.FashionMNIST(
  root="~/data",
  train=False,
  download=True,
  transform=transform,
  )
 
- # Create data loaders
+ # Create data loaders.
  train_dataloader = DataLoader(training_data, batch_size=batch_size)
  test_dataloader = DataLoader(test_data, batch_size=batch_size)
 
@@ -69,7 +69,7 @@ def train_func_per_worker(config: Dict):
  epochs = config["epochs"]
  batch_size = config["batch_size_per_worker"]
 
- # Get dataloaders inside the worker training function
+ # Get dataloaders inside worker training function
  train_dataloader, test_dataloader = get_dataloaders(batch_size=batch_size)
 
  # [1] Prepare Dataloader for distributed training
@@ -81,7 +81,7 @@ def train_func_per_worker(config: Dict):
  model = NeuralNetwork()
 
  # [2] Prepare and wrap your model with DistributedDataParallel
- # Move the model to the correct GPU/CPU device
+ # Move the model the correct GPU/CPU device
  # ============================================================
  model = ray.train.torch.prepare_model(model)
 
@@ -137,9 +137,9 @@ def train_fashion_mnist(num_workers=2, use_gpu=False):
  scaling_config=scaling_config,
  )
 
- # [4] Start distributed training
+ # [4] Start Distributed Training
  # Run `train_func_per_worker` on all workers
- # ==========================================
+ # =============================================
  result = trainer.fit()
  print(f"Training result: {result}")
 

python/ray/train/examples/transformers/transformers_torch_trainer_basic.py

@@ -14,8 +14,8 @@
 from ray.train.torch import TorchTrainer
 
 
-# [1] Define a training function that includes all your training logic
-# ====================================================================
+# [1] Define a training function that includes all your training logics
+# =====================================================================
 def train_func(config):
  # Datasets
  dataset = load_dataset("yelp_review_full")
@@ -34,15 +34,15 @@ def tokenize_function(examples):
  "bert-base-cased", num_labels=5
  )
 
- # Evaluation metrics
+ # Evaluation Metrics
  metric = evaluate.load("accuracy")
 
  def compute_metrics(eval_pred):
  logits, labels = eval_pred
  predictions = np.argmax(logits, axis=-1)
  return metric.compute(predictions=predictions, references=labels)
 
- # Hugging Face Trainer
+ # HuggingFace Trainer
  training_args = TrainingArguments(
  output_dir="test_trainer", evaluation_strategy="epoch", report_to="none"
  )
@@ -59,7 +59,7 @@ def compute_metrics(eval_pred):
  # ===============================================
  trainer.add_callback(RayTrainReportCallback())
 
- # [3] Prepare your trainer for Ray Data integration
+ # [3] Prepare your trainer for Ray Data Integration
  # =================================================
  trainer = prepare_trainer(trainer)
 

python/ray/train/torch/torch_trainer.py

@@ -23,9 +23,7 @@ class TorchTrainer(DataParallelTrainer):
  4. Runs the input ``train_loop_per_worker(train_loop_config)``
  on all workers.
 
- For more details, see :ref:`PyTorch Guide <train-pytorch>`,
- :ref:`PyTorch Lightning Guide <train-pytorch-lightning>`,
- or :ref:`Hugging Face Transformers Guide <train-pytorch-transformers>`.
+ For more details, see the :ref:`PyTorch User Guide <train-pytorch>`.
 
  Example: