• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 104 subscribers
  • Views 317 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

EDGE-AI-STUDIO: Can compile models with edge-AI tools, but not with EDGE-AI-STUDIO.

Liu Tsing-Yue
Intellectual 291 points
Part Number: EDGE-AI-STUDIO

Hi experts,

We're trying to test model inference time on EDGE-AI-STUDIO, using am62a.

We also make sure it is able to complie our onnx model using edge-AI tools.

The tag is 09_00_00_06, also using am62a.

However, we were unable to complie model from model analyzer in EDGE-AI-STUDIO.

It will show this error:

Fullscreen
1
2
3
---> 11 output = list(sess.run(None, {input_details[0].name : preprocess_for_multitask(sess, calib_images[num])}))[0]
RuntimeException: [ONNXRuntimeError] : 6 : RUNTIME_EXCEPTION : Non-zero status code returned while running TIDL_0 node. Name:'TIDLExecutionProvider_TIDL_0_0' Status Message: CHECK failed: (index) < (current_size_):
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

What might be the cause of this error?

Thanks.

  • 0
    Intellectual 291 points

    Also, we uploaded the complied model and try to test inference time, but only cpu is working and not c7x_mma.

    If we can get the result from c7x_mma, we won't need to complie models

  • 0 in reply to Liu Tsing-Yue
    Intellectual 291 points

    this is the log for compilation

    0488.custon-model-onnx_out.log
    Fullscreen
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    tidl_tools_path = /home/root/notebooks/tidl_tools
    artifacts_folder = custom-artifacts/onnx/multitask-v3_3_2
    tidl_tensor_bits = 8
    debug_level = 1
    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 = 1
    tidl_calibration_options:bias_calibration_iterations = 3
    mixed_precision_factor = -1.000000
    model_group_id = 0
    power_of_2_quantization = 2
    enable_high_resolution_optimization = 0
    pre_batchnorm_fold = 1
    add_data_convert_ops = 0
    XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

  • 0 in reply to Liu Tsing-Yue
    TI__Guru**** 450031 points

    I have brought this thread to attention of the EDGE AI Studio team. 

    Thanks

    ki

  • 0 in reply to Liu Tsing-Yue
    TI__Genius 9235 points

    Liu,
    Can you let me know which ONNX opset is your model exported for?

    Thanks,
    Varun