fix penn dataset (#1552)

parmeet · web-flow · commit 9561cdeb2015 · 2022-01-28T17:40:10.000-05:00
diff --git a/torchtext/datasets/penntreebank.py b/torchtext/datasets/penntreebank.py
@@ -47,7 +47,7 @@ def PennTreebank(root, split: Union[Tuple[str], str]):
         hash_dict={os.path.join(root, os.path.basename(URL[split])): MD5[split]},
         hash_type="md5",
     )
-    cache_dp = HttpReader(cache_dp).end_caching(mode="w", same_filepath_fn=True)
+    cache_dp = HttpReader(cache_dp).end_caching(mode="wb", same_filepath_fn=True)
     data_dp = FileOpener(cache_dp, mode="r")
     # remove single leading and trailing space from the dataset
     return data_dp.readlines(return_path=False).map(lambda t: t.strip())

Original file line number	Diff line number	Diff line change
`@@ -47,7 +47,7 @@ def PennTreebank(root, split: Union[Tuple[str], str]):`
`47`	`47`	`hash_dict={os.path.join(root, os.path.basename(URL[split])): MD5[split]},`
`48`	`48`	`hash_type="md5",`
`49`	`49`	`)`
`50`		`- cache_dp = HttpReader(cache_dp).end_caching(mode="w", same_filepath_fn=True)`
	`50`	`+ cache_dp = HttpReader(cache_dp).end_caching(mode="wb", same_filepath_fn=True)`
`51`	`51`	`data_dp = FileOpener(cache_dp, mode="r")`
`52`	`52`	`# remove single leading and trailing space from the dataset`
`53`	`53`	`return data_dp.readlines(return_path=False).map(lambda t: t.strip())`