backnotprop / fast-style-transfer-coreml Goto Github PK

So I changed the code to avoid the crash caused by trying to store every models in memory.

I get weird outputs :

• wave gives a green background
• rain_princess gives some kind of glitchy result

when I frozen pb file from 6.7M to 1.7M,then run tf_converter.conver, error info as :
Traceback (most recent call last):
File "/Users/icarus/Documents/icarusCode/Python/neural-style-transfer/fast-style-transfer-coreml-master/fast_style_transfer/convert_graph/convert.py", line 5, in
image_input_names = ['img_placeholder__0'])
File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/tfcoreml/_tf_coreml_converter.py", line 586, in convert
custom_conversion_functions=custom_conversion_functions)
File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/tfcoreml/_tf_coreml_converter.py", line 337, in _convert_pb_to_mlmodel
convert_ops_to_layers(context)
File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/tfcoreml/_ops_to_layers.py", line 178, in convert_ops_to_layers
translator(op, context)
File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/tfcoreml/_layers_common.py", line 77, in effectively_constant_op
add_const(context, out.name, x, out.name)
File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/tfcoreml/_layers_common.py", line 26, in add_const
_add_const(context, name, x, output_name, shape)
File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/tfcoreml/_shape_sensitive_layers.py", line 44, in _add_const
'Add constant: cannot add a constant in which all the dimensions '
AssertionError: Add constant: cannot add a constant in which all the dimensions (Seq, C, H, W) are of non-unit size

Import your CoreML and run iOS code, it has this issue: