diff options
Diffstat (limited to 'tensorflow/tools/graph_transforms/README.md')
-rw-r--r-- | tensorflow/tools/graph_transforms/README.md | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/tensorflow/tools/graph_transforms/README.md b/tensorflow/tools/graph_transforms/README.md index 36a5c01a0c..73af8699b2 100644 --- a/tensorflow/tools/graph_transforms/README.md +++ b/tensorflow/tools/graph_transforms/README.md @@ -179,7 +179,7 @@ bazel-bin/tensorflow/tools/graph_transforms/transform_graph \ --outputs='softmax:0' \ --transforms='\ strip_unused_nodes(type=float, shape="1,299,299,3") \ -fold_constants \ +fold_constants(ignore_errors=true) \ fold_batch_norms \ fold_old_batch_norms\ ' @@ -715,7 +715,7 @@ shape arguments let you control the attributes of any new Placeholders that are created. Plain `type` and `shape` set global defaults, but if you have different inputs with varying characteristics, you'll need to pass in a list of arguments where the preceding name specifies what layer each applies to. For example, if -you had two inputs in1 and in2, you could call `strip_unused_node(name=in1, +you had two inputs in1 and in2, you could call `strip_unused_nodes(name=in1, type_for_name=int32, shape_for_name="2,3", name=in2, type_for_name=float, shape_for_name="1,10,10,3")`. |