Adding source files

ja_training.py

@@ -82,12 +82,12 @@ def create_dataloaders(dataset_name, args):
     ds_kwargs = {"streaming":True, "chunksize":40<<20}
     #train_data = load_dataset('text', data_files={'train': ["wiki_mrph.txt"]},
     #    split="train[:90%]", **ds_kwargs)
-    train_data = load_dataset('text', data_files={'train': ["wiki_mrph_split_aa"]},
+    train_data = load_dataset('text', data_files={'train': ["../ja-test-data/wiki_mrph_split_aa"]},
         split='train', **ds_kwargs)
     print(train_data)
     #valid_data = load_dataset('text', data_files={'train': ["wiki_mrph.txt"]},
     #    split="train[-10%:]", **ds_kwargs)
-    valid_data = load_dataset('text', data_files={'train': ["wiki_mrph_split_ab"]},
+    valid_data = load_dataset('text', data_files={'train': ["../ja-test-data/wiki_mrph_split_ab"]},
         split='train', **ds_kwargs)
     print(valid_data)
     #train_data = chunked((x for x in dataset), 1000)