tensorflow1.14 onnx-tf 1.5.0 onnx 1.5.0
pytorch->onnx->tensorflow ,加載pb模型的時候出現如下錯誤
Invalid argument: No OpKernel was registered to support Op 'PyFunc' with these attrs. Registered devices: [CPU], Registered kernels:
<no registered kernels>
[[Node: PyFunc = PyFunc[Tin=[DT_FLOAT, DT_INT32, DT_INT32, DT_INT32, DT_INT32, DT_INT32, DT_STRING], Tout=[DT_FLOAT], token="pyfunc_0"](Relu, PyFunc/input_1, PyFunc/input_2, PyFunc/input_3, PyFunc/input_4, PyFunc/input_5, PyFunc/input_6)]]
參考https://github.com/onnx/onnx-tensorflow/issues/167
from onnx_tf.backend import prepare
model = onnx.load("duc_se.onnx")
tf_rep = prepare(model, strict=False)
測試alphapose的.pd模型和pytorch的結果是一致的