ArjunGoray1

  • Mar 23, 2024
  • Joined Jan 28, 2024
  • 0 best answers
  • In Error with blob converter

    Hey, I am trying to convert my SSD Mobilenet Tensorflow 2 object detection api model to the blob file using the blob converter, but nothing I do seems to be working. First, I converted the savedmodel to onnx, but when I convert the onnx to blob using the blob converter, I keep getting the following errors. Thanks in advance to anyone who can help!

    Command failed with exit code 1, command: /app/venvs/venv2022_1/bin/python /app/model_compiler/openvino_2022.1/converter.py --precisions FP16 --output_dir /tmp/blobconverter/9ab0a327fad04dba803c89b352a0b40a --download_dir /tmp/blobconverter/9ab0a327fad04dba803c89b352a0b40a --name model --model_root /tmp/blobconverter/9ab0a327fad04dba803c89b352a0b40a

    [ ERROR ] -------------------------------------------------
    [ ERROR ] ----------------- INTERNAL ERROR ----------------
    [ ERROR ] Unexpected exception happened.
    [ ERROR ] Please contact Model Optimizer developers and forward the following information:
    [ ERROR ] Check 'element_type.is_real()' failed at core/src/preprocess/preprocess_steps_impl.cpp:100:
    Mean preprocessing can be applied to 'float' inputs. Consider using of 'convert_element_type' before scaling. Current type is: u8

    [ ERROR ] Traceback (most recent call last):
    File "/app/venvs/venv2022_1/lib/python3.8/site-packages/openvino/tools/mo/main.py", line 533, in main
    ret_code = driver(argv)
    File "/app/venvs/venv2022_1/lib/python3.8/site-packages/openvino/tools/mo/main.py", line 493, in driver
    ret_res = moc_emit_ir(ngraph_function, argv)
    File "/app/venvs/venv2022_1/lib/python3.8/site-packages/openvino/tools/mo/moc_frontend/serialize.py", line 19, in moc_emit_ir
    apply_preprocessing(ov_function=ngraph_function, argv=argv)
    File "/app/venvs/venv2022_1/lib/python3.8/site-packages/openvino/tools/mo/back/preprocessing.py", line 417, in apply_preprocessing
    ov_function = prep.build()
    RuntimeError: Check 'element_type.is_real()' failed at core/src/preprocess/preprocess_steps_impl.cpp:100:
    Mean preprocessing can be applied to 'float' inputs. Consider using of 'convert_element_type' before scaling. Current type is: u8

    [ ERROR ] ---------------- END OF BUG REPORT --------------
    [ ERROR ] -------------------------------------------------

  • In Run Model on Host and Use Oak-D lite for fast depth coordinates

    I am trying to run yolov8n/s on my host (jetson nano) using the center rgb camera as the source for object detection, and then use the Oak-D Lite for the depth detection. I have tried to find code online for how to do so, but I could not find any. I also tried going through the luxonis documentation, but that did not help me much as I didn't know which functions to use. How would I go about doing this?

    Thanks in advance to anyone who can help!