|
我的模型输入[320000,1],在模型内部做reshape到(1600,200,1)后接vgg处理,这样转换之后pb模型可以正常使用,但rknn模型inference时候就报错,而且模型转换的log没有error或warning,rknn-toolkit版本是0.9.9。
E Catch exception when init model!
T Traceback (most recent call last):
T File "rknn/api/rknn_base.py", line 625, in rknn.api.rknn_base.RKNNBase.inference
T File "rknn/api/rknn_runtime.py", line 275, in rknn.api.rknn_runtime.RKNNRuntime.build_graph
T Exception: RKNN init failed. error code: RKNN_ERR_MODEL_INVALID
模型输入reshape代码
input_data = Input(name='the_input', shape=(self.INPUT_SIZE, 1), dtype='float32')
input_wavImage = Reshape((self.AUDIO_LENGTH, self.AUDIO_FEATURE_LENGTH, 1), name='Reshape')(input_data)
rknn模型转换代码
rknn.config()
# Load TensorFlow Model
print('--> Loading model')
rknn.load_tensorflow(tf_pb='./model.pb',
inputs=['the_input'],
outputs=['Activation0/truediv'],
input_size_list=[[320000, 1]]) |
|
|