整合数据处理和数据训练文件

data/data_process.py

@@ -103,7 +103,7 @@ def data_pre():
     data['test'] = get_data('test')
 
     with open(pickle_file, 'wb') as file:
-        print(data)
+        # print(data)
         pickle.dump(data, file)
 
     print('num_train: ' + str(len(data['train'])))

train.py

@@ -12,6 +12,7 @@ from models.LAS.seq2seq import Seq2Seq
 from data.data_load import AiShellDataset, pad_collate
 from config.conf import device, print_freq, vocab_size, num_workers, sos_id, eos_id
 from utils.util import get_logger, save_checkpoint, AverageMeter
+from data.data_process import data_pre
 
 parser = argparse.ArgumentParser(
     "End-to-End Automatic Speech Recognition Training "
@@ -226,4 +227,5 @@ def main():
 
 
 if __name__ == '__main__':
+    data_pre()
     main()