// TODO - only works for 4D tensor without batch axis
target_batch_idx = axes.index(0) // the assumed TF batch axis
target_height_idx = axes.index(1) // the assumed TF height axis
target_width_idx = axes.index(2) // the assumed TF width axis
target_channel_idx = axes.index(3) // the assumed TF channel axis
coreml_axes = [0]*4
coreml_axes[target_batch_idx] = 0
coreml_axes[target_height_idx] = 2
coreml_axes[target_width_idx] = 3
coreml_axes[target_channel_idx] = 1
context.builder.add_permute(output_name, coreml_axes, input_name, output_name)
context.translated[output_name] = True
After Change
//Permutation without swapping appears to give wrong results, which may be
//a bug in coreml itself
if axes[1] != 1 and axes[2] != 2 and axes[3] != 3:
assert False, "Only swapping permutes work"
// First, work out where the indicies should move in TF
target_idx = (axes.index(0),
axes.index(1),
axes.index(2),
axes.index(3))
// Translate from NHWC to NCHW order
target_idx = (target_idx[0],
target_idx[3],
target_idx[1],
target_idx[2])
def translate_transpose(idx):
if idx == 0:
