Update mypin-voice-dataset.py

mypin-voice-dataset.py

@@ -1,5 +1,6 @@
 import datasets
 import os
+import random
 
 class MyDataset(datasets.GeneratorBasedBuilder):
     def _info(self):
@@ -21,17 +22,43 @@ class MyDataset(datasets.GeneratorBasedBuilder):
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
-                gen_kwargs={"filepath": metadata, "audio_dir": data_dir},
+                gen_kwargs={
+                    "filepath": metadata, 
+                    "audio_dir": data_dir
+                    "split": "train",
+                    "split_ratio": 0.8,
+                },
             ),
+            datasets.SplitGenerator(
+                name=datasets.Split.VALIDATION,
+                gen_kwargs={
+                    "filepath": metadata, 
+                    "audio_dir": data_dir
+                    "split": "eval",
+                    "split_ratio": 0.2,
+                },
+            )
         ]
 
     def _generate_examples(self, filepath, audio_dir):
         import json
 
         with open(filepath, "r") as f:
-            for idx, line in enumerate(f):
-                record = json.loads(line)
-                yield idx, {
-                    "sentence": record["sentence"],
-                    "audio": os.path.join(audio_dir, record["file_name"]),  # Path to the audio file
-                }
+            data = [json.loads(line) for line in f]
+
+        # Shuffle data for randomness
+        random.shuffle(data)
+
+        # Split data into train and eval
+        split_index = int(len(data) * split_ratio)
+        if split == "train":
+            examples = data[:split_index]
+        else:
+            examples = data[split_index:]
+        
+        for idx, line in enumerate(f):
+            record = json.loads(line)
+            yield idx, {
+                "sentence": record["sentence"],
+                "audio": os.path.join(audio_dir, record["file_name"]),  # Path to the audio file
+            }