Comments

• Hi Willian To help diagnose your KL730 KNEO Pi boot issue after your firmware update attempt, could you please provide the following details: Firmware: Which part of the firmware did you update (e.g., U-Boot, kernel, rootfs, device tree)? What is th…

  in KNEO PI boot problem Comment by PowHuang July 2025
• 也有可能是USB線材的問題,要確定是usb 3的線 如果是720dongle用usb typec to A,可以試試 typec端正反面插,可能只有其中一面可以到super speed

  in 樹莓派連接720問題 Comment by PowHuang February 2022
• Hi Nagata a. host_lib(1.1.1 R2) only compatible with SDK_1.1.1 b.Only the release version host_lib_v1.0.0 support command "kl720_isi_load_model" and host_lib_v1.0.0 work with SDK_1.3 https://www.kneron.com/forum/uploads/975/D9YR9SJ68FRS.pn…

  in SDK v1.3 and host_lib Examples Comment by PowHuang February 2022
• Hi Gábor Simor I tested two opcv and all workable. 1.Follow document to build Opencv_347 and set Environment Variables. Install Dependency - Document Center (kneron.com) 2.download open_341 zip from this link. It don't need to build. huihut/OpenCV-M…

  in KL720 Dongle webcam example error Comment by PowHuang February 2022
• switch to UART mode is correct when run "python flash_programmer.py -a ... " you should get the message as picture: https://www.kneron.com/forum/uploads/990/DE632WUR61P1.png what error message you got?

  in Program Flashing Comment by PowHuang February 2022
• https://www.kneron.com/forum/uploads/622/6TKBCR92A6H4.png 正確install_libusb.sh後,也會建立/etc/udev/rules.d/10-local.rules檔案 就不須使用sudo來執行範例範例 python要確定有安裝以下套件 $ pip list KneronPLUS       1.2.1         KneronPLUS-YOLO    1.2.1…

  in KL720 dongle deploy on RK3399 problem Comment by PowHuang December 2021
• https://www.kneron.com/forum/uploads/886/LDIBBRAQ8ETJ.png 安裝libusb步驟有問題, "root is not in the sudoers file"網路可以找到權限問題討論

  in KL720 dongle deploy on RK3399 problem Comment by PowHuang December 2021
• -KL520有支援樹莓派 -目前還沒有公開的520 openpose 模型 -可以嘗試用kneron toolchain 對Lightweight OpenPose模型做轉換成kneron的格式 但模型有部分不支援的'operaters需要自行修改 GitHub - Daniil-Osokin/lightweight-human-pose-estimation.pytorch: Fast and ac…

  in OpenPose 應用 Comment by PowHuang November 2021
• 你是指這個? Kneron Academy 只支援KL520, 且如果KL520更新到KDP2 就無法用Kneron Academy Kneron Academy 無法連接上webcam ,這跟520/720無關, 預設是使用cam "0" ,請確認你的webcam資源有沒有被占用 https://www.kneron.com/forum/uploads/010/22T220TB53…

  in KL720 在win10安裝完 Dependency 和 KDP2 Firmware , 但是 Kneron Academy 無法連接上webcam Comment by PowHuang October 2021
• 我執行KL720-cam_isi_pre_post_host_fcos並沒有問題 你可能要確認你的環境 或試試先做一次update_fw 把\app_binaries\KL720\solution_companion裡的2個bin檔copy 到\app_binaries\KL720\dfu\ready_to_load >py main.py -t KL720-update_fw https://www.kneron.com/…

  in How to use python script on KL720 dongle to inference Comment by PowHuang October 2021
• host_lib_v1.0.0 python examples_kl720 內沒有跑模型的範例 520 / 720用法不同 請下載GitHub - kneron/host_lib: This project contains python examples for the Kneron USB dongle 這個版本host_lib python範例較完整

  in How to use python script on KL720 dongle to inference Comment by PowHuang October 2021
• Hi Jasmine, tensorflow example mnst是黑白的模型 所以"img_channel"要改"L"

  in 如果要用其他模型轉換的onnx檔，該如何設定input_params.json? Comment by PowHuang October 2021
• 是的, 可在Quantization步驟放100張跟 training / inference相關的圖片, 準確度會比較高

  in 提升準確度之方法詢問 Comment by PowHuang September 2021
• In the KL720-yolo_public_example , input image using RGB565 color format (2 bytes per pixel). The input data is similar with orginal picture but not same. So this example will get diferent output.

  in Inference code for testing docker toolchain example Comment by PowHuang August 2021
• Try "pip install Pillow"

  in 推論時有錯誤 Comment by PowHuang August 2021
• Currently , no speech example. In constants.py .the ModelType are Face detect / object detect / OCR purpose, all for image. If you have a speech model . You need create your data pre process and post process function.

  in Is it possible to write inference code for speech models using KDP Wrapper ? Comment by PowHuang August 2021
• My model_id setting is 32768. km = ktc.ModelConfig(32768, "0001", "720", onnx_path="/workspace/examples/LittleNet/LittleNet.onnx") your model_id is 1001 So you need change model_id setting to 1001 in yolo_public_example…

  in Inference code for testing docker toolchain example Comment by PowHuang August 2021
• could you upload your littlenet models_720.nef

  in Inference code for testing docker toolchain example Comment by PowHuang August 2021
• In KL520, ISI mean load model from flash. dme mean load model from usb. But KL720 remove dme api, and use kdp_isi_config to set model location. https://www.kneron.com/forum/uploads/695/Y1R9DLLH8XDM.jpg param : CONFIG_USE_FLASH_MODEL = bit(0)   # 1…

  in Whats the difference ISI and DME mode and which mode should I use for inferencing Littlenet model ? Comment by PowHuang August 2021
• The python examples is good start for learn APIs for config / inference / get result.

  in Whats the difference ISI and DME mode and which mode should I use for inferencing Littlenet model ? Comment by PowHuang August 2021
• I post a comment for littlenet test at the 720 forunm . The littlenet is just a example for toolchain model convert. Yolo example is a more better to learn kneron APIs. kdp_wrapper is necessary for python example 720 not support dme mode now, 720 us…

  in Questions regarding testing /workspace/examples/LittleNet/LittleNet.onnx model on KL720 Comment by PowHuang August 2021
• The easy way to test : use the https://github.com/kneron/host_lib KL720-yolo_public_example I attach the files as zip file. Abdullah_0001_112_112.bin is the input image in RGB565 format. D:\Kneron\host_lib\host_lib-master\python>py main.py -t…

  in Inference code for testing docker toolchain example Comment by PowHuang August 2021
• 圖片是用 Netron 模型視覺化工具,可圖示了解模型結構

  in 如果要用其他模型轉換的onnx檔，該如何設定input_params.json? Comment by PowHuang August 2021
• inpur_params.json 應修改"model_input_name": "inputs", https://www.kneron.com/forum/uploads/804/1CPO4FS3TT0J.jpg

  in 如果要用其他模型轉換的onnx檔，該如何設定input_params.json? Comment by PowHuang August 2021
• Same as another case. ktc.kneron_inference 720 bie, must add platform=720 inf_results = ktc.kneron_inference(input_data, bie_file="/data1/output.bie", platform=720, input_names=["data_out"])

  in When I put bie path in ktc.kneron_inference instead of onnx path it gives segmentation error Comment by PowHuang August 2021
• Could you share your kl720_image_inference.py file?

  in When I put bie path in ktc.kneron_inference instead of onnx path it gives segmentation error Comment by PowHuang August 2021
• ONNX is float , Kneron NPU is INT8 Kneron toolchain make model compression to int8 type (NEF). It have some lossy in inference.

  in I get two different values for inf_results when I give input model as ONNX and NEF MODEL Comment by PowHuang August 2021
• If you use Vmware , try add CPU and RAM to setting https://www.kneron.com/forum/uploads/825/5DOEEFLHAHQQ.jpg

  in When I put bie path in ktc.kneron_inference instead of onnx path it gives segmentation error Comment by PowHuang August 2021
• For KL720 BIE simulator need extra parameter "platform=720" fixed_results = ktc.kneron_inference(input_data, bie_file=bie_path, input_names=["data_out"], platform=720, radix=radix) https://www.kneron.com/forum/uploads/965/RCTQT1L…

  in When In try to create nef file using LittleNet.onnx model for kl720 , it fails Comment by PowHuang August 2021
• Ubuntu need install opencv too, there is steps for reference git clone https://github.com/opencv/opencv.git cd opencv git checkout 6d4f66472e14b29b8e1623859cfebfdc67f677c3 mkdir build cd build cmake -D CMAKE_INSTALL_PREFIX=/usr/local -D CMAKE_BUILD…

  in How to view uart logs in kl720 Comment by PowHuang August 2021
More Comments