🐛 Fix wrong cache dir in speech_recognition_dataset.py

hezarai · Feb 7, 2024 · d7dadd6 · d7dadd6
1 parent 0e59778
commit d7dadd6
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/hezar/data/datasets/speech_recognition_dataset.py b/hezar/data/datasets/speech_recognition_dataset.py
@@ -2,7 +2,7 @@
 
 from dataclasses import dataclass
 
-from datasets import Audio, load_dataset, load_from_disk
+from datasets import Audio, load_dataset
 
 from .dataset import Dataset
 from ..data_collators import SpeechRecognitionDataCollator
@@ -51,7 +51,7 @@ def __init__(self, config: SpeechRecognitionDatasetConfig, split=None, **kwargs)
         )
 
     def _load(self, split):
-        data = load_dataset(self.config.path, split=split)
+        data = load_dataset(self.config.path, split=split, cache_dir=self.cache_dir)
         data = data.cast_column(self.config.audio_column, Audio(sampling_rate=self.config.sampling_rate))
         return data