MLM training fails with no validation file(same as #12406 for pytorch now) (#12517)

* Validation split percentage to be used for custom data files also

Issue same as https://github.com/huggingface/transformers/issues/12406 fixed for pytorch branch run_mlm.py

* Validation split added in the right place

* Update run_clm.py

* validation split added for custom files

* Validation split added for custom files

* Update run_plm.py

* fixed validation split for custom files as input for pytorch examples in lm

* Update run_clm_no_trainer.py

* args modified

examples/pytorch/language-modeling/run_clm.py

@@ -278,6 +278,21 @@ def main():
         if extension == "txt":
             extension = "text"
         raw_datasets = load_dataset(extension, data_files=data_files, cache_dir=model_args.cache_dir)
+        # If no validation data is there, validation_split_percentage will be used to divide the dataset.
+        if "validation" not in raw_datasets.keys():
+            raw_datasets["validation"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[:{data_args.validation_split_percentage}%]",
+                cache_dir=model_args.cache_dir,
+            )
+            raw_datasets["train"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[{data_args.validation_split_percentage}%:]",
+                cache_dir=model_args.cache_dir,
+            )
+
     # See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
     # https://huggingface.co/docs/datasets/loading_datasets.html.
 

examples/pytorch/language-modeling/run_clm_no_trainer.py

@@ -253,6 +253,19 @@ def main():
         if extension == "txt":
             extension = "text"
         raw_datasets = load_dataset(extension, data_files=data_files)
+        # If no validation data is there, validation_split_percentage will be used to divide the dataset.
+        if "validation" not in raw_datasets.keys():
+            raw_datasets["validation"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[:{args.validation_split_percentage}%]",
+            )
+            raw_datasets["train"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[{args.validation_split_percentage}%:]",
+            )
+
     # See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
     # https://huggingface.co/docs/datasets/loading_datasets.html.
 

examples/pytorch/language-modeling/run_mlm.py

@@ -285,6 +285,22 @@ def main():
         if extension == "txt":
             extension = "text"
         raw_datasets = load_dataset(extension, data_files=data_files, cache_dir=model_args.cache_dir)
+
+        # If no validation data is there, validation_split_percentage will be used to divide the dataset.
+        if "validation" not in raw_datasets.keys():
+            raw_datasets["validation"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[:{data_args.validation_split_percentage}%]",
+                cache_dir=model_args.cache_dir,
+            )
+            raw_datasets["train"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[{data_args.validation_split_percentage}%:]",
+                cache_dir=model_args.cache_dir,
+            )
+
     # See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
     # https://huggingface.co/docs/datasets/loading_datasets.html.
 

examples/pytorch/language-modeling/run_mlm_no_trainer.py

@@ -265,6 +265,19 @@ def main():
         if extension == "txt":
             extension = "text"
         raw_datasets = load_dataset(extension, data_files=data_files)
+        # If no validation data is there, validation_split_percentage will be used to divide the dataset.
+        if "validation" not in raw_datasets.keys():
+            raw_datasets["validation"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[:{args.validation_split_percentage}%]",
+            )
+            raw_datasets["train"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[{args.validation_split_percentage}%:]",
+            )
+
     # See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
     # https://huggingface.co/docs/datasets/loading_datasets.html.
 

examples/pytorch/language-modeling/run_plm.py

@@ -280,6 +280,21 @@ def main():
         if extension == "txt":
             extension = "text"
         raw_datasets = load_dataset(extension, data_files=data_files, cache_dir=model_args.cache_dir)
+        # If no validation data is there, validation_split_percentage will be used to divide the dataset.
+        if "validation" not in raw_datasets.keys():
+            raw_datasets["validation"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[:{data_args.validation_split_percentage}%]",
+                cache_dir=model_args.cache_dir,
+            )
+            raw_datasets["train"] = load_dataset(
+                extension,
+                data_files=data_files,
+                split=f"train[{data_args.validation_split_percentage}%:]",
+                cache_dir=model_args.cache_dir,
+            )
+
     # See more about loading any type of standard or custom dataset (from files, python dict, pandas DataFrame, etc) at
     # https://huggingface.co/docs/datasets/loading_datasets.html.