is_train rename

verl/trainer/main_ppo.py

@@ -188,8 +188,8 @@ def run(self, config):
         from verl.utils.dataset.rl_dataset import collate_fn
 
         # Create training and validation datasets.
-        train_dataset = create_rl_dataset(config.data.train_files, config.data, tokenizer, processor, train=True)
-        val_dataset = create_rl_dataset(config.data.val_files, config.data, tokenizer, processor, train=False)
+        train_dataset = create_rl_dataset(config.data.train_files, config.data, tokenizer, processor, is_train=True)
+        val_dataset = create_rl_dataset(config.data.val_files, config.data, tokenizer, processor, is_train=False)
         train_sampler = create_rl_sampler(config.data, train_dataset)
 
         # Initialize the PPO trainer.
@@ -214,7 +214,7 @@ def run(self, config):
         trainer.fit()
 
 
-def create_rl_dataset(data_paths, data_config, tokenizer, processor, train=True):
+def create_rl_dataset(data_paths, data_config, tokenizer, processor, is_train=True):
     """Create a dataset.
 
     Arguments:
@@ -243,8 +243,8 @@ def create_rl_dataset(data_paths, data_config, tokenizer, processor, train=True)
                 f"The custom dataset class '{data_config.custom_cls.name}' from "
                 f"'{data_config.custom_cls.path}' must inherit from torch.utils.data.Dataset"
             )
-    elif "datagen" in data_config and data_config.datagen.get("path", None) is not None and train:
-        # If a data generation strategy is specified, use the DataGenDataset class
+    elif "datagen" in data_config and data_config.datagen.get("path", None) is not None and is_train:
+        # If a data generation strategy is specified, use the DynamicGenDataset class
         from verl.utils.dataset.dynamicgen_dataset import DynamicGenDataset
 
         dataset_cls = DynamicGenDataset