PB 模型转换RKNN 问题

[复制链接] · 发表于 2021-6-30 19:24:20

python3 test.py
--> Loading model
W:tensorflow:From /home/tf/anaconda3/envs/rk1808_TFV1/lib/python3.6/site-packages/rknn/api/rknn.py:75: The name tf.GraphDef is deprecated. Please use tf.compat.v1.GraphDef instead.

W:tensorflow:From /home/tf/anaconda3/envs/rk1808_TFV1/lib/python3.6/site-packages/rknn/api/rknn.py:75: extract_sub_graph (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.compat.v1.graph_util.extract_sub_graph`
W:tensorflow:From /home/tf/anaconda3/envs/rk1808_TFV1/lib/python3.6/site-packages/rknn/api/rknn.py:75: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.

W:tensorflow:From /home/tf/anaconda3/envs/rk1808_TFV1/lib/python3.6/site-packages/rknn/api/rknn.py:75: The name tf.global_variables_initializer is deprecated. Please use tf.compat.v1.global_variables_initializer instead.

E Unknow layer "tensorarraygatherv3"
E Try match TensorArrayGatherV3 Postprocessor/BatchMultiClassNonMaxSuppression/map/TensorArrayStack/TensorArrayGatherV3 failed, catch exception!
E Catch exception when loading tensorflow model: 2_28.pb!
E Traceback (most recent call last):
E File "rknn/base/RKNNlib/converter/convert_tf.py", line 750, in rknn.base.RKNNlib.converter.convert_tf.convert_tf.match_paragraph_and_param
E File "rknn/base/RKNNlib/RKNNnet.py", line 211, in rknn.base.RKNNlib.RKNNnet.RKNNNet.new_layer
E File "rknn/api/rknn_log.py", line 312, in rknn.api.rknn_log.RKNNLog.e
E ValueError: Unknow layer "tensorarraygatherv3"
E During handling of the above exception, another exception occurred:
E Traceback (most recent call last):
E File "rknn/api/rknn_base.py", line 209, in rknn.api.rknn_base.RKNNBase.load_tensorflow
E File "rknn/base/RKNNlib/RK_nn.py", line 47, in rknn.base.RKNNlib.RK_nn.RKnn.load_tensorflow
E File "rknn/base/RKNNlib/app/importer/import_tensorflow.py", line 232, in rknn.base.RKNNlib.app.importer.import_tensorflow.Importensorflow.run
E File "rknn/base/RKNNlib/converter/convert_tf.py", line 849, in rknn.base.RKNNlib.converter.convert_tf.convert_tf.match_paragraph_and_param
E File "rknn/api/rknn_log.py", line 312, in rknn.api.rknn_log.RKNNLog.e
E ValueError: Try match TensorArrayGatherV3 Postprocessor/BatchMultiClassNonMaxSuppression/map/TensorArrayStack/TensorArrayGatherV3 failed, catch exception!
done
--> Building model
Traceback (most recent call last):
File "test.py", line 33, in <module>
rknn.build(do_quantization=False)
File "/home/tf/anaconda3/envs/rk1808_TFV1/lib/python3.6/site-packages/rknn/api/rknn.py", line 228, in build
inputs = self.rknn_base.net.get_input_layers()
AttributeError: 'NoneType' object has no attribute 'get_input_layers'

cat test.py
from rknn.api import RKNN

INPUT_SIZE = 64

if __name__ == '__main__':
# 创建RKNN执行对象
rknn = RKNN()
# 配置模型输入，用于NPU对数据输入的预处理
# channel_mean_value='0 0 0 255'，那么模型推理时，将会对RGB数据做如下转换
# (R - 0)/255, (G - 0)/255, (B - 0)/255。推理时，RKNN模型会自动做均值和归一化处理
# reorder_channel=’0 1 2’用于指定是否调整图像通道顺序，设置成0 1 2即按输入的图像通道顺序不做调整
# reorder_channel=’2 1 0’表示交换0和2通道，如果输入是RGB，将会被调整为BGR。如果是BGR将会被调整为RGB
#图像通道顺序不做调整
#rknn.config(channel_mean_value='0 0 0 255', reorder_channel='0 1 2')
rknn.config(reorder_channel='0 1 2')

# 加载TensorFlow模型
# tf_pb='digital_gesture.pb'指定待转换的TensorFlow模型
# inputs指定模型中的输入节点
# outputs指定模型中输出节点
# input_size_list指定模型输入的大小
print('--> Loading model')
rknn.load_tensorflow(tf_pb='2_28.pb',
inputs=['image_tensor'],
outputs=['detection_boxes', 'detection_scores', 'detection_multiclass_scores', 'detection_classes', 'num_detections', 'raw_detection_boxes', 'raw_detection_scores'],
input_size_list=[[800, 600, 3]])
print('done')

# 创建解析pb模型
#指定不进行量化,量化会减小模型的体积和提升运算速度，但是会有精度的丢失
#do_quantization=False
print('--> Building model')
rknn.build(do_quantization=False)
print('done')

# 导出保存rknn模型文件
rknn.export_rknn('./digital_gesture.rknn')

# Release RKNN Context
rknn.release()

想问下这是哪里问题？tensorarraygatherv3 Layer 不支持问题

只看该作者 · 发表于 2021-7-1 10:30:06

斑竹能不能帮忙看下