load_onnx报错

[复制链接] · 发表于 2019-5-15 09:36:46

rknn 1.0.0和0.9.9都试过。

模型文件：https://s3.amazonaws.com/onnx-model-zoo/arcface/resnet100.onnx

转换程序：

from rknn.api import RKNN



rknn = RKNN()



rknn.config(channel_mean_value='127.5 127.5 127.5 128', reorder_channel='0 1 2')



print('--> Loading model')

rknn.load_onnx(model='./resnet100.onnx')

print('done')



print('--> Building model')

rknn.build(do_quantization=False)

print('done')



rknn.export_rknn('./resnet100.rknn')



rknn.release()
复制代码

报错:

--> Loading model

W Not match tensor MatMul_464:out0 

E Catch exception when loading onnx model: ./resnet100.onnx!

T Traceback (most recent call last):

T   File "rknn/api/rknn_base.py", line 344, in rknn.api.rknn_base.RKNNBase.load_onnx

T   File "rknn/base/RKNNlib/converter/convert_onnx.py", line 1037, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx.match_paragraph_and_param

T   File "rknn/base/RKNNlib/converter/convert_onnx.py", line 950, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx._onnx_push_ready_tensor

T TypeError: 'NoneType' object is not iterable

复制代码

只看该作者 · 发表于 2019-5-17 12:24:28

rknn = RKNN(verbose=True)
请把verbose打开，不然看不到什么有用的 log

只看该作者 · 发表于 2019-5-17 14:20:31

目前rknn版本1.0.0

W verbose file path is invalid, debug info will not dump to file.

--> Loading model

I Current ONNX Model use ir_version 3 opset_version 8

D import clients finished

I build output layer attach_BatchNormalization_466:out0

I Try match BatchNormalization_466:out0

I Match r_bn_v6 [['BatchNormalization_466', 'Initializer_770', 'Initializer_771', 'Initializer_772', 'Initializer_773']] [['BatchNormalization', 'Constant_0', 'Constant_1', 'Constant_2', 'Constant_3']] to [['batchnormalize']]

I Try match Add_465:out0

I Match r_add [['Add_465']] [['Add']] to [['add']]

I Try match MatMul_464:out0

W Not match tensor MatMul_464:out0 

E Catch exception when loading onnx model: ./resnet100.onnx!

T Traceback (most recent call last):

T   File "rknn/api/rknn_base.py", line 344, in rknn.api.rknn_base.RKNNBase.load_onnx

T   File "rknn/base/RKNNlib/converter/convert_onnx.py", line 1037, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx.match_paragraph_and_param

T   File "rknn/base/RKNNlib/converter/convert_onnx.py", line 950, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx._onnx_push_ready_tensor

T TypeError: 'NoneType' object is not iterable

done

--> Building model

E Model or data is None, please load model first.

done

E RKNN model data is None, please load model first!

复制代码

只看该作者 · 发表于 2019-5-17 14:23:40

感谢版版还记得这个帖子，现在onnx版本1.5.0，我降到过1.2.0，也是有问题

只看该作者 · 发表于 2019-6-24 14:55:33

请问如何解决 load_onnx 报错.

T TypeError: 'NoneType' object is not iterable

使用rknn v1.0.0
load_tflite OK

完整LOG

--> config model

done

--> Loading ONNX model

I Current ONNX Model use ir_version 4 opset_version 9

D import clients finished

I build output layer attach_BatchNormalization_31:out0

I build output layer attach_Div_39:out0

I Try match Div_39:out0

I Match r_div [['Div_39']] [['Div']] to [['divide']]

I Try match BatchNormalization_31:out0

I Match r_bn_v6 [['BatchNormalization_31', 'Initializer_83', 'Initializer_79', 'Initializer_81', 'Initializer_82']] [['BatchNormalization', 'Constant_0', 'Constant_1', 'Constant_2', 'Constant_3']] to [['batchnormalize']]

I Try match BatchNormalization_36:out0

I Match r_bn_v6 [['BatchNormalization_36', 'Initializer_69', 'Initializer_65', 'Initializer_67', 'Initializer_68']] [['BatchNormalization', 'Constant_0', 'Constant_1', 'Constant_2', 'Constant_3']] to [['batchnormalize']]

I Try match Unsqueeze_38:out0

I Match r_unsqueeze [['Unsqueeze_38']] [['Unsqueeze']] to [['reshape']]

I Try match Conv_30:out0

I Match r_conv [['Conv_30', 'Initializer_78', 'Initializer_77']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]

I Try match Conv_35:out0

I Match r_conv [['Conv_35', 'Initializer_64', 'Initializer_63']] [['Conv', 'Constant_0', 'Constant_1']] to [['convolution']]

I Try match ReduceL2_37:out0

W Not match tensor ReduceL2_37:out0 

E Catch exception when loading onnx model: /N/share/build/rknn/rknn-toolkit/example/mobilenet_v1_onnx/superpoint_v1_ws.onnx!

T Traceback (most recent call last):

T   File "rknn/api/rknn_base.py", line 344, in rknn.api.rknn_base.RKNNBase.load_onnx

T   File "rknn/base/RKNNlib/converter/convert_onnx.py", line 1037, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx.match_paragraph_and_param

T   File "rknn/base/RKNNlib/converter/convert_onnx.py", line 950, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx._onnx_push_ready_tensor

T TypeError: 'NoneType' object is not iterable

Load superpoint_v1.onnx failed!



Process finished with exit code 255
复制代码

只看该作者 · 发表于 2019-6-24 14:57:58

yaowei 发表于 2019-5-17 14:20
目前rknn版本1.0.0

Hi @jefferyzhang， @yaowei，先谢谢指点！

只看该作者 · 发表于 2020-6-30 09:01:06

yaowei 发表于 2019-5-17 14:20
目前rknn版本1.0.0

您好，请问W verbose file path is invalid, debug info will not dump to file.这个报错你是怎么解决的？

只看该作者 · 发表于 2020-6-30 11:23:32

wschong 发表于 2019-6-24 14:57
Hi @jefferyzhang， @yaowei，先谢谢指点！

rknn toolkit可以升级到1.3.2比较好定位问题。

只看该作者 · 发表于 2020-6-30 11:23:40

wschong 发表于 2019-6-24 14:57
Hi @jefferyzhang， @yaowei，先谢谢指点！

rknn toolkit可以升级到1.3.2比较好定位问题。