|
|
问题背景：
我的目的是将线性判别分析算法的程序rknn化，将程序运行在香橙派5B的NPU上。
线性判别分析算法的rknn化，实现步骤是：
①用pytorch库一步步参照sklearn库中的线性判别算法的实现过程写出线性判别分析的pytorch程序；
②将pytorch程序写出的线性判别分析算法导出为pt模型；
③将pt模型转换为onnx模型；
④将onnx模型转换为rknn模型；
但是将rknn模型运行起来报错；
运行rknn模型的程序如下：
from rknnlite.api import RKNNLite as RKNN
import numpy as np
import pandas as pd
from sklearn.preprocessing import StandardScaler
import pickle
# model='linear_model.rknn'
# model='CNN_Lee20231201.rknn'
model='LDA_Lee20240122.rknn'
# input_test_data = [23, 32, 197, 1022, 1, 1, 18, 24, 169, 458, 1, 1, 25, 31, 173, 697, 1, 1, 27, 36, 195, 1221, 1, 1]
# 定义特征名称和数据
feature_names = ['AverageAbsoluteValue_Channel1', 'RootMeanSquare_Channel1', 'MeanFrequency_Channel1' \
, 'Hjorth Activity_Channel1', 'Hjorth Mobility_Channel1', 'Hjorth Complexity_Channel1' \
, 'AverageAbsoluteValue_Channel2', 'RootMeanSquare_Channel2', 'MeanFrequency_Channel2' \
, 'Hjorth Activity_Channel2', 'Hjorth Mobility_Channel2', 'Hjorth Complexity_Channel2' \
, 'AverageAbsoluteValue_Channel3', 'RootMeanSquare_Channel3', 'MeanFrequency_Channel3' \
, 'Hjorth Activity_Channel3', 'Hjorth Mobility_Channel3', 'Hjorth Complexity_Channel3' \
, 'AverageAbsoluteValue_Channel4', 'RootMeanSquare_Channel4', 'MeanFrequency_Channel4' \
, 'Hjorth Activity_Channel4', 'Hjorth Mobility_Channel4', 'Hjorth Complexity_Channel4']
features_Sample = [23, 32, 197, 1022, 1, 1, 18, 24, 169, 458, 1, 1, 25, 31, 173, 697, 1, 1, 27, 36, 195, 1221, 1, 1]
features_MergedArray = np.array(features_Sample)
features_MergedArrayD2 = features_MergedArray.reshape(1, -1)
# 转换为DataFrame并设置特征名称
features_MergedDataFrame = pd.DataFrame(features_MergedArrayD2, columns=feature_names)
with open('scaler_PytorchCNNVersion.pkl', 'rb') as f:
scaler = pickle.load(f)
features_scaled = scaler.transform(features_MergedDataFrame)
# 将 float64 转换为 float32
#features_scaled_float32 = features_scaled.astype(np.float32)
features_scaled_float32 = features_scaled.astype(np.float16)
# rknn=RKNN()
rknn=RKNN(verbose=True)
ret=rknn.load_rknn(path=model)
print('--> Init runtime environment')
ret = rknn.init_runtime(
#target=None,
#target="rk3588",
# target_sub_class=None,
device_id=None,
# perf_debug=False,
# eval_mem=False,
async_mode=False,
core_mask=RKNN.NPU_CORE_AUTO
#core_mask=0
)
if ret != 0:
print('Init runtime environment failed')
exit(ret)
print('done')
print(22222222111111222222222)
# output=rknn.inference(inputs=[features_scaled_float32])
output=rknn.inference(inputs=[features_scaled_float32])
print(111111)
print(output)
print(111111)
报错如下：
D RKNN: [15:10:01.805] normalize target: CPU
D RKNN: [15:10:01.805] Get NPU frequency: 1000MHz
W RKNN: [15:10:01.805] failed to get DDR freq, take 1560MHz as default freq
D RKNN: [15:10:01.805] Get DDR frequency: 1560MHz
D RKNN: [15:10:01.805] dump npy tensor to: dumps/000_InputOperator_0_shape_1_24_float.npy
D RKNN: [15:10:01.806] dump npy tensor to: dumps/001_Reshape_Gemm_3_2conv_transpose1_output_shape_1_24_1_1_float.npy
D RKNN: [15:10:01.806] dump npy tensor to: dumps/002_Conv_4_conv_shape_1_1_1_1_float.npy
D RKNN: [15:10:01.806] dump npy tensor to: dumps/003_Reshape_6_shape_1_float.npy
E RKNN: [15:10:01.805] failed open dump path: dumps/000_InputOperator_0_shape_1_24_float.npy
E RKNN: [15:10:01.806] failed open dump path: dumps/001_Reshape_Gemm_3_2conv_transpose1_output_shape_1_24_1_1_float.npy
E RKNN: [15:10:01.806] failed open dump path: dumps/002_Conv_4_conv_shape_1_1_1_1_float.npy
E RKNN: [15:10:01.806] failed open dump path: dumps/003_Reshape_6_shape_1_float.npy
E RKNN: [15:10:01.807] DoElementwiseLogical unknown inputA layout:0
/tmp/tmp7b2e49h0：行 3: 4631 已放弃 （核心已转储） python /home/orangepi/Pycharm/pycharm-community-2023.1/plugins/python-ce/helpers/pydev/pydevd.py --multiprocess --qt-support=auto --client 127.0.0.1 --port 38573 --file /home/orangepi/EMG_RKNN_Lee/myRKNN_Project/02_inference/inference_LDA_rknn_Lee20240124.py
ERROR conda.cli.main_run:execute(49): `conda run python /home/orangepi/Pycharm/pycharm-community-2023.1/plugins/python-ce/helpers/pydev/pydevd.py --multiprocess --qt-support=auto --client 127.0.0.1 --port 38573 --file /home/orangepi/EMG_RKNN_Lee/myRKNN_Project/02_inference/inference_LDA_rknn_Lee20240124.py` failed. (See above for error)
经过断点调试，发现报错的语句为：
output=rknn.inference(inputs=[features_scaled_float32])
|
|
收藏
闽公网安备 35010202001061号