Init runtime environment failed

固件：Debian10rknn-toolkit :1.2.1
rknn-api:1.2.0
drv:0.9.9
tensorflow:1.14.0(之前因为path的问题重新刷写了一遍固件，刷写完自带的这个版本。。。Wiki里的1.13.x试过无法安装，使用pip3只有1.14.0和2.0. 另外官方是默认带中科大的源吗？我之前自己换过一次源。。但是重新刷写完自带就油中科大的源）
使用的官方的DeeplabV3的pb模型，论坛里有人成果转换过（指明了输入和输出节点）并推理http://t.rock-chips.com/forum.php?mod=viewthread&tid=227&highlight=Deeplab，我按照同样的方法也成功转化了，但是推理报错，信息如下

1. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/ceil.py:10: The name tf.ceil is deprecated. Please use tf.math.ceil instead.
   
2. 
   
3. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/depth_to_space.py:12: The name tf.depth_to_space is deprecated. Please use tf.compat.v1.depth_to_space instead.
   
4. 
   
5. W:tensorflow:
   
6. The TensorFlow contrib module will not be included in TensorFlow 2.0.
   
7. For more information, please see:
   
8.   * https://github.com/tensorflow/community/blob/master/rfcs/20180907-contrib-sunset.md
   
9.   * https://github.com/tensorflow/addons
   
10.   * https://github.com/tensorflow/io (for I/O related ops)
    
11. If you depend on functionality not listed there, please file an issue.
    
12. 
    
13. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/log.py:10: The name tf.log is deprecated. Please use tf.math.log instead.
    
14. 
    
15. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/random_normal.py:9: The name tf.random_normal is deprecated. Please use tf.random.normal instead.
    
16. 
    
17. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/random_uniform.py:9: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.
    
18. 
    
19. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/upsample.py:13: The name tf.image.resize_images is deprecated. Please use tf.image.resize instead.
    
20. 
    
21. /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/common/__init__.py:87: UserWarning: FrontendHandler.get_outputs_names is deprecated. It will be removed in future release.. Use node.outputs instead.
    
22.   warnings.warn(message)
    
23. -->loading model
    
24. loading model done
    
25. --> Init runtime environment
    
26. E RKNNAPI: rknn_init,  recv(MsgLoadAck) fail, -9(ERROR_PIPE) != 368!
    
27. E Catch exception when init runtime!
    
28. E Traceback (most recent call last):
    
29. E   File "rknn/api/rknn_base.py", line 788, in rknn.api.rknn_base.RKNNBase.init_runtime
    
30. E   File "rknn/api/rknn_runtime.py", line 303, in rknn.api.rknn_runtime.RKNNRuntime.build_graph
    
31. E Exception: RKNN init failed. error code: RKNN_ERR_DEVICE_UNAVAILABLE
    
32. Init runtime environment failed
    

复制代码PB模型下载地址http://download.tensorflow.org/m ... g_2018_01_29.tar.gz

rknn-api:1.2.0
drv:0.9.9
drv版本和api版本都不匹配，能正常运行才怪。。。
先把DRV和api都升级到1.3.0再说。

另外我们Toybrick的板子从来没有默认中科大源。。。

jefferyzhang 发表于 2020-5-7 20:56
rknn-api:1.2.0
drv:0.9.9
drv版本和api版本都不匹配，能正常运行才怪。。。

版主您好，按您说的我已经全部升级到了1.3版本
rknn-toolkit：1.3.2
rknn-api:1.3.2
rknn-drv:1.3.1
(我有点疑惑，其中有一个组件RKNN-RK3399RPO,安装您的安装提示，这个应该是api...，提示版本1.3.0-3，但运行示例demo显示的api是1.3.2）
再次转化在Debian10正式版固件条件下可以转化的PB模型，提示以下错误

1. --> Loading model
   
2. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/ceil.py:10: The name tf.ceil is deprecated. Please use tf.math.ceil instead.
   
3. 
   
4. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/depth_to_space.py:12: The name tf.depth_to_space is deprecated. Please use tf.compat.v1.depth_to_space instead.
   
5. 
   
6. W:tensorflow:
   
7. The TensorFlow contrib module will not be included in TensorFlow 2.0.
   
8. For more information, please see:
   
9.   * https://github.com/tensorflow/community/blob/master/rfcs/20180907-contrib-sunset.md
   
10.   * https://github.com/tensorflow/addons
    
11.   * https://github.com/tensorflow/io (for I/O related ops)
    
12. If you depend on functionality not listed there, please file an issue.
    
13. 
    
14. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/log.py:10: The name tf.log is deprecated. Please use tf.math.log instead.
    
15. 
    
16. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/random_normal.py:9: The name tf.random_normal is deprecated. Please use tf.random.normal instead.
    
17. 
    
18. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/random_uniform.py:9: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.
    
19. 
    
20. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/handlers/backend/upsample.py:13: The name tf.image.resize_images is deprecated. Please use tf.image.resize instead.
    
21. 
    
22. /home/toybrick/.local/lib/python3.7/site-packages/onnx_tf/common/__init__.py:87: UserWarning: FrontendHandler.get_outputs_names is deprecated. It will be removed in future release.. Use node.outputs instead.
    
23.   warnings.warn(message)
    
24. W:tensorflow:From /home/toybrick/.local/lib/python3.7/site-packages/rknn/api/rknn.py:67: extract_sub_graph (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
    
25. Instructions for updating:
    
26. Use `tf.compat.v1.graph_util.extract_sub_graph`
    
27. E Catch exception when loading tensorflow model: ./frozen_inference_graph.pb!
    
28. E Traceback (most recent call last):
    
29. E   File "rknn/api/rknn_base.py", line 215, in rknn.api.rknn_base.RKNNBase.load_tensorflow
    
30. E   File "rknn/base/RKNNlib/converter/convert_tf.py", line 527, in rknn.base.RKNNlib.converter.convert_tf.convert_tf.pre_process
    
31. E   File "rknn/base/RKNNlib/converter/tensorflowloader.py", line 96, in rknn.base.RKNNlib.converter.tensorflowloader.TF_Graph_Preprocess.pre_proces
    
32. E   File "rknn/base/RKNNlib/converter/tensorflowloader.py", line 555, in rknn.base.RKNNlib.converter.tensorflowloader.TF_Graph_Preprocess.import_tf_ruler_module
    
33. E   File "/home/toybrick/.local/lib/python3.7/site-packages/rknn/base/RKNNlib/converter/tfruler/__init__.py", line 2, in <module>
    
34. E     from .tf_ruler_generate import r_add_n_template, r_concatv2_template, r_split_template
    
35. E ImportError: cannot import name 'r_add_n_template' from 'tfruler.tf_ruler_generate' (/home/toybrick/.local/lib/python3.7/site-packages/rknn/base/RKNNlib/converter/tfruler/tf_ruler_generate.cpython-37m-aarch64-linux-gnu.so)
    
36. done
    
37. --> Building model
    
38. Traceback (most recent call last):
    
39.   File "deeplab3_rknn.py", line 31, in <module>
    
40.     rknn.build(do_quantization=False)
    
41.   File "/home/toybrick/.local/lib/python3.7/site-packages/rknn/api/rknn.py", line 222, in build
    
42.     inputs = self.rknn_base.net.get_input_layers()
    
43. AttributeError: 'NoneType' object has no attribute 'get_input_layers'
    

复制代码
这个模型是论坛里有人转化成功的，看时间应该是1.3.0固件发布之前，他也成功推理了。我现在无法转化（可以推理，但是我不能确定是否可以用1.2.1转化的模型在1.3版本进行推理），现在的rknn模型是之前版本转化的，我记得有文档说过之前转化的模型可能不通用，因此想使用1.3版本转化，输入输出层也是按之前他提供的方式定义的，之前无法推理应该就是您说的DRV版本过低，我现在是想如果1.3版本无法完成，我退回到1.2版本只更新DRV来处理
        最后源的问题是这样的，我板子第一次刷写完官方Debian10固件后我自己更换过源，但是我重新刷写固件后，这个之前更换的源依然存在起效。。。。

tensorflow 你要用1.13.x的，不能用1.14，因为有些op14版本加了v2识别不了。
按我经验get_input_layers出错都是你转换时候问题。你转换完本地仿真先跑下看看能不能推理

jefferyzhang 发表于 2020-5-8 13:15
tensorflow 你要用1.13.x的，不能用1.14，因为有些op14版本加了v2识别不了。
按我经验get_input_layers出错 ...

感谢版主回答。。。问题是rock-chips/pypi/simple只有1.10.1和1.14.0以及2.0.0（我看到您在别的帖子里说到不建议使用1.13一下版本)。Wiki教程里的百度云版本也是1.10.1版本，如何找到aarch64 版本的1.13版本呢。。。。

86667780 发表于 2020-5-8 13:30
感谢版主回答。。。问题是rock-chips/pypi/simple只有1.10.1和1.14.0以及2.0.0（我看到您在别的帖子里说 ...

啥？你转换模型也在板子上做的？？？
额，我建议你现在pc上试试，仿真只有pc才有