Fix character with wrong encodding in documentation (pytorch#12761)

kennyhorror · facebook-github-bot · commit e027f7a913f1 · 2018-10-17T10:20:45.000-07:00
Summary: Pull Request resolved: pytorch#12761 , is not really , and thus it can fail some of the Python 2 import. Reviewed By: weiyangfb Differential Revision: D10423231 fbshipit-source-id: 3738c0b9d2f52aa47eef06250f84c5933a38783f
diff --git a/torch/_torch_docs.py b/torch/_torch_docs.py
@@ -4243,7 +4243,7 @@ def parse_kwargs(desc):
     #   3. SparseTensor._values().shape = (nnz, SparseTensor.shape[sparse_dim:])
     #
     # For instance, to create an empty sparse tensor with nnz = 0, dense_dim = 0 and
-    # sparse_dim = 1 (hence indices is a 2D tensor of shape = (1‚ 0))
+    # sparse_dim = 1 (hence indices is a 2D tensor of shape = (1, 0))
     >>> S = torch.sparse_coo_tensor(torch.empty([1, 0]), [], [1])
     tensor(indices=tensor([], size=(1, 0)),
            values=tensor([], size=(0,)),

Original file line number	Diff line number	Diff line change
`@@ -4243,7 +4243,7 @@ def parse_kwargs(desc):`
`4243`	`4243`	`# 3. SparseTensor._values().shape = (nnz, SparseTensor.shape[sparse_dim:])`
`4244`	`4244`	`#`
`4245`	`4245`	`# For instance, to create an empty sparse tensor with nnz = 0, dense_dim = 0 and`
`4246`		`- # sparse_dim = 1 (hence indices is a 2D tensor of shape = (1‚ 0))`
	`4246`	`+ # sparse_dim = 1 (hence indices is a 2D tensor of shape = (1, 0))`
`4247`	`4247`	`>>> S = torch.sparse_coo_tensor(torch.empty([1, 0]), [], [1])`
`4248`	`4248`	`tensor(indices=tensor([], size=(1, 0)),`
`4249`	`4249`	`values=tensor([], size=(0,)),`