提交 157526f3 authored 作者: Frederic's avatar Frederic

fix dtype problem in a test.

上级 44a208cf
......@@ -172,10 +172,10 @@ class test_ifelse(unittest.TestCase, utt.TestOptimizationMixin):
p = [p0, p1, p2, p3]
# in my code these vars are the result of applying scan
ften0 = tensor.ftensor3('ft0')
fmat1 = tensor.fmatrix('fm1')
ften2 = tensor.ftensor3('ft2')
fmat3 = tensor.fmatrix('fm3')
ften0 = tensor.tensor3('ft0', dtype=self.dtype)
fmat1 = tensor.matrix('fm1', dtype=self.dtype)
ften2 = tensor.tensor3('ft2', dtype=self.dtype)
fmat3 = tensor.matrix('fm3', dtype=self.dtype)
# then I keep only the last iteration
fsub0 = ften0[-1]
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论