• State Resolved
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 101 subscribers
  • Views 150 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

PROCESSOR-SDK-J722S: [TIDL] Convert default onnx/tflite model to TIDL format fail

ken zhoung
Prodigy 100 points
Part Number: PROCESSOR-SDK-J722S
Other Parts Discussed in Thread: AM69A

Tool/software:

Hi team:

TIDL version : 10.00.08.00

SDK LINUX image: 10.00.00.08

I use the example code: ~/edgeai-tidl-tools/examples/jupyter_notebooks/custom-model-onnx.ipynb and custom-model-tflite.ipynb to convert default model(onnx/tflite) file to TIDL format,

where default model is provided from /edgeai-tidl-tools/examples/jupyter_notebooks/models/public

But I get the two error below:

-------- Running Calibration in Float Mode to Collect Tensor Statistics --------
[=============================================================================] 100 %

------------------ Fixed-point Calibration Iteration [1 / 3]: ------------------
Illegal instruction (core dumped)                                             ] 25 %
[TIDL Import]  ERROR: Failed to run calibration pass, system command returned error: 33792 -- [tidl_import_core.cpp, 675]
[TIDL Import]  ERROR: Failed to run Calibration - Failed in function: tidlRunQuantStatsTool -- [tidl_import_core.cpp, 1720]
[TIDL Import] [QUANTIZATION] ERROR:  - Failed in function: TIDL_quantStatsFixedOrFloat -- [tidl_import_quantize.cpp, 3574]
[TIDL Import] [QUANTIZATION] ERROR:  - Failed in function: TIDL_runIterativeCalibration -- [tidl_import_quantize.cpp, 3722]
[TIDL Import] [QUANTIZATION] ERROR:  - Failed in function: TIDL_import_quantize -- [tidl_import_quantize.cpp, 4545]
[TIDL Import]  ERROR:  - Failed in function: TIDL_import_backend -- [tidl_import_core.cpp, 4139]
[TIDL Import]  ERROR:  - Failed in function: TIDL_runtimesPostProcessNet -- [tidl_runtimes_import_common.cpp, 1390]
tidl_tfLiteRtImport_delegate.cpp Invoke 647 
100%|█████████████████████████████████████████████████████████████████████████████████████████████████| 4/4 [00:10<00:00,  2.56s/it]

******** WARNING ******* : Could not open custom-artifacts/tflite/mobilev1//allowedNode.txt for reading... Entire model will run on ARM without any delegation to TIDL !

Please help me to check this symptom,

Thanks for your kindly help.

  • 0
    TI__Expert 6700 points

    Hi, 

    What specific model are you using?

    Chris

  • 0 in reply to Chris Tsongas
    Prodigy 100 points

    Hi Chris:

    I use the ~/edgeai-tidl-tools/examples/jupyter_notebooks/models/public/onnx/resnet18_opset9.onnx 

    and  ~/edgeai-tidl-tools/examples/jupyter_notebooks/models/public/tflite/mobilenet_v1_1.0_224.onnx 

    above model are the example model in TIDL(10.00.08.00)

    I would like to confirm if the absence of a GPU in my testing environment could affect the results of the process?

    Thanks

  • 0 in reply to ken zhoung
    Prodigy 100 points

    Hi Chris,

    I occasionally encounter a "Bus error (core dumped)" in TIDL_subgraphRtCreate (log below) while running the model resnet18_opset9.onnx. I have already checked, and the available memory appears to be sufficient.

     

    [Ken debug] show support providers : ['TIDLExecutionProvider', 'TIDLCompilationProvider', 'CPUExecutionProvider']
[Debug] TIDL_TOOLS_PATH = /home/ken/TIDL00.00.08.00/do_once/edgeai-tidl-tools/tidl_tools
2024-12-11 21:31:15.243468669 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer4.1.bn1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243501699 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer4.0.bn2.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243510738 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer3.0.downsample.1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243517742 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer4.0.downsample.1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243524477 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer3.0.bn1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243531038 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer3.1.bn2.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243538926 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer2.0.downsample.1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243545474 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer3.1.bn1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243551952 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer1.0.bn2.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243558777 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer1.1.bn2.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243565775 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'bn1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243573424 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer1.0.bn1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243580562 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer2.1.bn1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243586884 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer1.1.bn1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243593299 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer4.1.bn2.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243600069 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer4.0.bn1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243606536 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer2.0.bn1.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243612767 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer3.0.bn2.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243620848 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer2.0.bn2.num_batches_tracked'. It is not used by any node and should be removed from the model.
2024-12-11 21:31:15.243628754 [W:onnxruntime:, graph.cc:3480 CleanUnusedInitializersAndNodeArgs] Removing initializer 'layer2.1.bn2.num_batches_tracked'. It is not used by any node and should be removed from the model.
tidl_tools_path                                 = /home/ken/TIDL00.00.08.00/do_once/edgeai-tidl-tools/tidl_tools 
artifacts_folder                                = custom-artifacts/onnx/resnet18_opset9.onnx 
tidl_tensor_bits                                = 8 
debug_level                                     = 3 
num_tidl_subgraphs                              = 16 
tidl_denylist                                   = 
tidl_denylist_layer_name                        = 
tidl_denylist_layer_type                        = 
tidl_allowlist_layer_name                       = 
model_type                                      =  
tidl_calibration_accuracy_level                 = 7 
tidl_calibration_options:num_frames_calibration = 4 
tidl_calibration_options:bias_calibration_iterations = 3 
mixed_precision_factor = -1.000000 
model_group_id = 0 
power_of_2_quantization                         = 2 
ONNX QDQ Enabled                                = 0 
enable_high_resolution_optimization             = 0 
pre_batchnorm_fold                              = 1 
add_data_convert_ops                            = 1 
output_feature_16bit_names_list                 =  
m_params_16bit_names_list                       =  
m_single_core_layers_names_list                 =  
Inference mode                                  = 0 
Number of cores                                 = 1 
reserved_compile_constraints_flag               = 1601 
partial_init_during_compile                     = 0 
ti_internal_reserved_1                          = 

========================= [Model Compilation Started] =========================

Model compilation will perform the following stages:
1. Parsing
2. Graph Optimization
3. Quantization & Calibration
4. Memory Planning

============================== [Version Summary] ==============================

-------------------------------------------------------------------------------
|          TIDL Tools Version          |              10_00_08_00             |
-------------------------------------------------------------------------------
|         C7x Firmware Version         |              10_00_02_00             |
-------------------------------------------------------------------------------
|            Runtime Version           |            1.14.0+10000005           |
-------------------------------------------------------------------------------
|          Model Opset Version         |                   9                  |
-------------------------------------------------------------------------------

NOTE: The runtime version here specifies ONNXRT_VERSION+TIDL_VERSION
Ex: 1.14.0+1000XXXX -> ONNXRT 1.14.0 and a TIDL_VERSION 10.00.XX.XX

============================== [Parsing Started] ==============================

[TIDL Import] [PARSER] WARNING: Network not identified as Object Detection network : (1) Ignore if network is not Object Detection network (2) If network is Object Detection network, please specify "model_type":"OD" as part of OSRT compilation options
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - MaxPool,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Add,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Add,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Add,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Add,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Add,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Add,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Add,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Conv,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Add,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Relu,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - GlobalAveragePool,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Flatten,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]
[TIDL Import] [PARSER] SUPPORTED: Layers type supported by TIDL --- layer type - Gemm,  Node name -  -- [tidl_onnxRtImport_core.cpp, 529]

------------------------- Subgraph Information Summary -------------------------
-------------------------------------------------------------------------------
|          Core           |      No. of Nodes       |   Number of Subgraphs   |
-------------------------------------------------------------------------------
| C7x                     |                      49 |                       1 |
| CPU                     |                       0 |                       x |
-------------------------------------------------------------------------------
Running Runtimes GraphViz - /home/ken/TIDL00.00.08.00/do_once/edgeai-tidl-tools/tidl_tools/tidl_graphVisualiser_runtimes.out custom-artifacts/onnx/resnet18_opset9.onnx/allowedNode.txt custom-artifacts/onnx/resnet18_opset9.onnx/tempDir/graphvizInfo.txt custom-artifacts/onnx/resnet18_opset9.onnx/tempDir/runtimes_visualization.svg 
============================= [Parsing Completed] =============================

TIDL_createStateImportFunc Started:
Compute on node : TIDLExecutionProvider_TIDL_0_0
  0,            Conv, 3, 1, input.1, 124
  1,            Relu, 1, 1, 124, 125
  2,         MaxPool, 1, 1, 125, 126
  3,            Conv, 3, 1, 126, 128
  4,            Relu, 1, 1, 128, 129
  5,            Conv, 3, 1, 129, 131
  6,             Add, 2, 1, 131, 132
  7,            Relu, 1, 1, 132, 133
  8,            Conv, 3, 1, 133, 135
  9,            Relu, 1, 1, 135, 136
 10,            Conv, 3, 1, 136, 138
 11,             Add, 2, 1, 138, 139
 12,            Relu, 1, 1, 139, 140
 13,            Conv, 3, 1, 140, 142
 14,            Relu, 1, 1, 142, 143
 15,            Conv, 3, 1, 143, 145
 16,            Conv, 3, 1, 140, 147
 17,             Add, 2, 1, 145, 148
 18,            Relu, 1, 1, 148, 149
 19,            Conv, 3, 1, 149, 151
 20,            Relu, 1, 1, 151, 152
 21,            Conv, 3, 1, 152, 154
 22,             Add, 2, 1, 154, 155
 23,            Relu, 1, 1, 155, 156
 24,            Conv, 3, 1, 156, 158
 25,            Relu, 1, 1, 158, 159
 26,            Conv, 3, 1, 159, 161
 27,            Conv, 3, 1, 156, 163
 28,             Add, 2, 1, 161, 164
 29,            Relu, 1, 1, 164, 165
 30,            Conv, 3, 1, 165, 167
 31,            Relu, 1, 1, 167, 168
 32,            Conv, 3, 1, 168, 170
 33,             Add, 2, 1, 170, 171
 34,            Relu, 1, 1, 171, 172
 35,            Conv, 3, 1, 172, 174
 36,            Relu, 1, 1, 174, 175
 37,            Conv, 3, 1, 175, 177
 38,            Conv, 3, 1, 172, 179
 39,             Add, 2, 1, 177, 180
 40,            Relu, 1, 1, 180, 181
 41,            Conv, 3, 1, 181, 183
 42,            Relu, 1, 1, 183, 184
 43,            Conv, 3, 1, 184, 186
 44,             Add, 2, 1, 186, 187
 45,            Relu, 1, 1, 187, 188
 46, GlobalAveragePool, 1, 1, 188, 189
 47,         Flatten, 1, 1, 189, 190
 48,            Gemm, 3, 1, 190, 191

Input tensor name -  input.1 
Output tensor name - 191 
  0%|                                                                                                                                                                                                                    | 0/4 [00:00<?, ?it/s]In TIDL_onnxRtImportInit subgraph_name=subgraph_0
Layer 0, subgraph id subgraph_0, name=191
Layer 1, subgraph id subgraph_0, name=input.1
==================== [Optimization for subgraph_0 Started] ====================

In TIDL_runtimesOptimizeNet: LayerIndex = 51, dataIndex = 50 
----------------------------- Optimization Summary -----------------------------
---------------------------------------------------------------------------------
|          Layer         | Nodes before optimization | Nodes after optimization |
---------------------------------------------------------------------------------
| TIDL_InnerProductLayer |                         1 |                        1 |
| TIDL_EltWiseLayer      |                         8 |                        8 |
| TIDL_ReLULayer         |                        17 |                        0 |
| TIDL_FlattenLayer      |                         1 |                        0 |
| TIDL_ConvolutionLayer  |                        20 |                       20 |
| TIDL_PoolingLayer      |                         2 |                        2 |
---------------------------------------------------------------------------------

=================== [Optimization for subgraph_0 Completed] ===================

In TIDL_runtimesPostProcessNet 
************ in TIDL_subgraphRtCreate ************ 
 The soft limit is 10240
The hard limit is 10240
MEM: Init ... !!!
MEM: Init ... Done !!!
 0.0s:  VX_ZONE_INIT:Enabled
 0.4s:  VX_ZONE_ERROR:Enabled
 0.5s:  VX_ZONE_WARNING:Enabled
 0.2289s:  VX_ZONE_INIT:[tivxInit:190] Initialization Done !!!
Bus error (core dumped)

    Do you have any suggestions or insights regarding this issue?

    Thank you for your kind assistance!

    Best regards,
    Ken 

  • +1 in reply to ken zhoung
    TI__Expert 6700 points

    Hi Ken,

    If you are getting intermittent failures, this is indicative of a problem in edgeai-tidl-tools/tidl_tools.  Something was probably not installed correctly, or the ENV var was not set correctly.  To re-install tidl_tools:

    Move the tidl_tools/ directory to another name.

    Make sure SOC is set, for example "export SOC=am69a"

    After that is installed make sure TIDL_TOOLS_PATH is set correctly. 
    TIDL_TOOLS_PATH=<SOME_PATH>/edgeai-tidl-tools/tidl_tools

    Also, can you try importing the model using OSRT-python?

    To do this, go to edgeai-tidl-tools/examples/osrt_python/ort and edit ../models_config.py to add your model (just copy a similar model config and set model_path to your model).  For example:

    model_path = os.path.join('/<path_to_your_model>/', 'mobilenet_v2_toy_model.onnx')

    then in edgeai-tidl-tools/examples/osrt_python/ort run:

    python3 ./onnxrt_ep.py -c -m <your_model_name>

    your_model_name will be specified as the first entry for your model in model_configs.py.  For example:

    "your_model_name": create_model_config(
    source=AttrDict(
    #model_url= 'file://home/a0194920local/10.0/edgeai-tidl-tools/models/public/mymodel.onnx',
    opt= True,
    infer_shape = True
    ), #set to dummy url for local file
    preprocess=AttrDict(
    resize=448,
    crop=224,
    data_layout="NCHW",
    resize_with_pad=False,
    reverse_channels=False,
    ),
    session=AttrDict(
    session_name='onnxrt' ,
    model_path = os.path.join('/home/a0194920local/', 'mobilenet_v2_toy_model.onnx'),
    #model_path = os.path.join(models_base_path, 'mymodel.onnx'),
    input_mean=[123.675, 116.28, 103.53], #confirmed by HQV
    input_scale = [0.017125, 0.017507, 0.017429], # confirmed by HQV
    ),
    task_type="classification",
    optional_options=AttrDict(
    #debug_level= 3,
    tensor_bits= 8,
    # "deny_list:layer_name" : '/original_model/backbone/arms.0/Mul',
    ),
    extra_info=AttrDict(
    num_images=numImages,
    num_classes=3,
    #model_type="detection"
    #model_type= 'segmentation',
    ),
    ),

  • 0 in reply to Chris Tsongas
    Prodigy 100 points

    Hi Chirs:

    I Try the python3 ./onnxrt_ep.py with default model, then can work.

    I will try my custom model.

    Sorry that I have a question : 

    If I aim to produce a converted model with an input type of uint8, is it mandatory for the original model to also have an input type of uint8?

    Alternatively, can I provide an original model with an input type of float and utilize the conversion tool to transform it into a model with an input type of uint8?

    Thanks for your kindly help.

  • 0 in reply to ken zhoung
    TI__Mastermind 38562 points

    The E2E responses will be delayed due to the TI Holiday (Christmas). Thank you for your patience.

  • 0 in reply to Praveen Rao
    TI__Expert 6700 points

    Hi Ken,

    Please open a new thread with your new question.  These threads are searchable, and we like to keep them to a single topic.

    Chris