$OpDocClipClipByAverageNorm
$OpDocClipClipByAverageNorm
Input tensor.
0-D (scalar) tensor > 0, specifying the maximum clipping value.
Name prefix for created ops.
Created op output.
$OpDocClipClipByGlobalNorm
$OpDocClipClipByGlobalNorm
Input tensors.
0-D (scalar) tensor > 0, specifying the maximum clipping value.
0-D (scalar) tensor containing the global norm to use. If not provided, globalNorm()
is used
to compute the norm.
Name prefix for created ops.
Tuple containing the clipped tensors as well as the global norm that was used for the clipping.
$OpDocClipClipByNorm
$OpDocClipClipByNorm
Input tensor.
0-D (scalar) tensor > 0, specifying the maximum clipping value.
1-D (vector) INT32
tensor containing the dimensions to use for computing the l2-norm. If
null
(the default), all dimensions are used.
Name prefix for created ops.
Created op output.
$OpDocClipClipByValue
$OpDocClipClipByValue
Input tensor.
0-D (scalar) tensor, or a tensor with the same shape as input
, specifying the minimum value
to clip by.
0-D (scalar) tensor, or a tensor with the same shape as input
, specifying the maximum value
to clip by.
Name prefix for created ops.
Created op output.
$OpDocClipGlobalNorm
$OpDocClipGlobalNorm
Input tensors.
Name prefix for created ops.
Created op output.