在进行模型转换时报错，多次尝试修改模型输入依然报错
oot@6a7cd5cbc4f3:/open_explorer/samples/ai_toolchain/horizon_model_convert_sample/03_classification/13_dms/mapper# ./01_check.sh
++ dirname ./01_check.sh
+ cd .
+ model_type=onnx
+ onnx_model=./dms.onnx
+ march=bayes-e
+ hb_mapper checker --model-type onnx --model ./dms.onnx --march bayes-e
2024-12-27 16:58:39,155 INFO log will be stored in /open_explorer/samples/ai_toolchain/horizon_model_convert_sample/03_classification/13_dms/mapper/hb_mapper_check
er.log
2024-12-27 16:58:39,155 INFO Start hb_mapper....
2024-12-27 16:58:39,155 INFO hbdk version 3.49.15
2024-12-27 16:58:39,155 INFO horizon_nn version 1.1.0
2024-12-27 16:58:39,155 INFO hb_mapper version 1.24.3
2024-12-27 16:58:39,169 INFO Model type: onnx
2024-12-27 16:58:39,169 INFO input names []
2024-12-27 16:58:39,169 INFO input shapes {}
2024-12-27 16:58:39,169 INFO Begin model checking....
2024-12-27 16:58:39,180 INFO Start to Horizon NN Model Convert.
2024-12-27 16:58:39,302 INFO Loading horizon_nn debug methods:set()
2024-12-27 16:58:39,302 INFO The activation calibration parameters:
calibration_type: fixed
2024-12-27 16:58:39,302 INFO The specified model compilation architecture: bayes-e.
2024-12-27 16:58:39,303 INFO The specified model compilation optimization parameters: [].
2024-12-27 16:58:39,303 INFO Start to prepare the onnx model.
2024-12-27 16:58:39,350 INFO Input ONNX Model Information:
ONNX IR version: 6
Opset version: ['ai.onnx v11', 'horizon v1']
Producer: pytorch v2.5.1
Domain: None
Version: None
Graph input:
...eshape_0: shape=[1, 1, 960, 1440], dtype=FLOAT32
...eshape_1: shape=[1, 3, 1, 1], dtype=FLOAT32
Graph output:
1487: shape=[1, 596], dtype=FLOAT32
2024-12-27 16:58:39,542 INFO End to prepare the onnx model.
2024-12-27 16:58:39,576 INFO Saving model to: ./.hb_check/original_float_model.onnx.
2024-12-27 16:58:39,576 INFO Start to optimize the onnx model.
2024-12-27 16:58:40,181 ERROR *** ERROR-OCCUR-DURING {horizon_nn.build_onnx} ***, error message: upported input data type:
The error model has been saved as replace_unsupported_averagepool_with_conv_pass_fail.onnx
2024-12-27 16:58:40,181 INFO End to optimize the onnx model.
2024-12-27 16:58:40,181 INFO End to Horizon NN Model Convert.
root@6a7cd5cbc4f3:/open_explorer/samples/ai_toolchain/horizon_model_convert_sample/03_classification/13_dms/mapper# ./01_check.sh
++ dirname ./01_check.sh
+ cd .
+ model_type=onnx
+ onnx_model=./dms.onnx
+ march=bayes-e
+ hb_mapper checker --model-type onnx --model ./dms.onnx --march bayes-e
2024-12-27 17:10:55,125 INFO log will be stored in /open_explorer/samples/ai_toolchain/horizon_model_convert_sample/03_classification/13_dms/mapper/hb_mapper_check
er.log
2024-12-27 17:10:55,125 INFO Start hb_mapper....
2024-12-27 17:10:55,125 INFO hbdk version 3.49.15
2024-12-27 17:10:55,125 INFO horizon_nn version 1.1.0
2024-12-27 17:10:55,125 INFO hb_mapper version 1.24.3
2024-12-27 17:10:55,138 INFO Model type: onnx
2024-12-27 17:10:55,139 INFO input names []
2024-12-27 17:10:55,139 INFO input shapes {}
2024-12-27 17:10:55,139 INFO Begin model checking....
2024-12-27 17:10:55,149 INFO Start to Horizon NN Model Convert.
2024-12-27 17:10:55,268 INFO Loading horizon_nn debug methods:set()
2024-12-27 17:10:55,269 INFO The activation calibration parameters:
calibration_type: fixed
2024-12-27 17:10:55,269 INFO The specified model compilation architecture: bayes-e.
2024-12-27 17:10:55,269 INFO The specified model compilation optimization parameters: [].
2024-12-27 17:10:55,269 INFO Start to prepare the onnx model.
2024-12-27 17:10:55,301 INFO Input ONNX Model Information:
ONNX IR version: 6
Opset version: ['ai.onnx v11', 'horizon v1']
Producer: pytorch v2.5.1
Domain: None
Version: None
Graph input:
...eshape_0: shape=[1, 1, 960, 1440], dtype=FLOAT32
Graph output:
1467: shape=[1, 596], dtype=FLOAT32
2024-12-27 17:10:55,440 INFO End to prepare the onnx model.
2024-12-27 17:10:55,475 INFO Saving model to: ./.hb_check/original_float_model.onnx.
2024-12-27 17:10:55,475 INFO Start to optimize the onnx model.
2024-12-27 17:10:56,244 ERROR *** ERROR-OCCUR-DURING {horizon_nn.build_onnx} ***, error message: upported input data type:
The error model has been saved as replace_unsupported_averagepool_with_conv_pass_fail.onnx
2024-12-27 17:10:56,244 INFO End to optimize the onnx model.
2024-12-27 17:10:56,244 INFO End to Horizon NN Model Convert.
root@6a7cd5cbc4f3:/open_explorer/samples/ai_toolchain/horizon_model_convert_sample/03_classification/13_dms/mapper# ./01_check.sh
++ dirname ./01_check.sh
+ cd .
+ model_type=onnx
+ onnx_model=./dms.onnx
+ march=bayes-e
+ hb_mapper checker --model-type onnx --model ./dms.onnx --march bayes-e
2024-12-27 17:12:01,809 INFO log will be stored in /open_explorer/samples/ai_toolchain/horizon_model_convert_sample/03_classification/13_dms/mapper/hb_mapper_check
er.log
2024-12-27 17:12:01,810 INFO Start hb_mapper....
2024-12-27 17:12:01,810 INFO hbdk version 3.49.15
2024-12-27 17:12:01,810 INFO horizon_nn version 1.1.0
2024-12-27 17:12:01,810 INFO hb_mapper version 1.24.3
2024-12-27 17:12:01,822 INFO Model type: onnx
2024-12-27 17:12:01,822 INFO input names []
2024-12-27 17:12:01,822 INFO input shapes {}
2024-12-27 17:12:01,822 INFO Begin model checking....
2024-12-27 17:12:01,833 INFO Start to Horizon NN Model Convert.
2024-12-27 17:12:01,948 INFO Loading horizon_nn debug methods:set()
2024-12-27 17:12:01,949 INFO The activation calibration parameters:
calibration_type: fixed
2024-12-27 17:12:01,949 INFO The specified model compilation architecture: bayes-e.
2024-12-27 17:12:01,949 INFO The specified model compilation optimization parameters: [].
2024-12-27 17:12:01,949 INFO Start to prepare the onnx model.
2024-12-27 17:12:01,981 INFO Input ONNX Model Information:
ONNX IR version: 6
Opset version: ['ai.onnx v11', 'horizon v1']
Producer: pytorch v2.5.1
Domain: None
Version: None
Graph input:
...eshape_0: shape=[1, 3, 320, 1440], dtype=FLOAT32
Graph output:
1467: shape=[1, 596], dtype=FLOAT32
2024-12-27 17:12:02,175 INFO End to prepare the onnx model.
2024-12-27 17:12:02,210 INFO Saving model to: ./.hb_check/original_float_model.onnx.
2024-12-27 17:12:02,210 INFO Start to optimize the onnx model.
2024-12-27 17:12:02,787 ERROR *** ERROR-OCCUR-DURING {horizon_nn.build_onnx} ***, error message: upported input data type:
The error model has been saved as replace_unsupported_averagepool_with_conv_pass_fail.onnx
2024-12-27 17:12:02,787 INFO End to optimize the onnx model.
2024-12-27 17:12:02,787 INFO End to Horizon NN Model Convert.
root@6a7cd5cbc4f3:/open_explorer/samples/ai_toolchain/horizon_model_convert_sample/03_classification/13_dms/mapper#