提交 d6c32c52 authored 作者: Dustin Webb's avatar Dustin Webb

Added clarifying comment.

Incremeneted GpuDnnSoftmax version and retested as an extra sanity test of my recent refactoring.
上级 bc6e4d4e
...@@ -540,7 +540,7 @@ err%(name)s = cudnnSoftmaxForward( ...@@ -540,7 +540,7 @@ err%(name)s = cudnnSoftmaxForward(
algo=algo, fail=sub['fail'], id=sub['struct_id'], name=name) algo=algo, fail=sub['fail'], id=sub['struct_id'], name=name)
def c_code_cache_version(self): def c_code_cache_version(self):
return (0, 1) return (0, 2)
@local_optimizer([GpuSoftmax]) @local_optimizer([GpuSoftmax])
......
...@@ -295,8 +295,8 @@ class test_SoftMax(unittest.TestCase): ...@@ -295,8 +295,8 @@ class test_SoftMax(unittest.TestCase):
check_types check_types
) )
# cuDNN cannot handle these test cases but the Theano softmax can so we # cuDNN R1 cannot handle these test cases but the Theano softmax can so
# test them only for the Theano softmax. # we test them only for the Theano softmax.
self._cmp(2 << 15, 5, f, f_gpu) self._cmp(2 << 15, 5, f, f_gpu)
self._cmp(0, 10, f, f_gpu) self._cmp(0, 10, f, f_gpu)
......
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