PowHuang
PowHuang
About
- Username
- PowHuang
- Joined
- Visits
- 375
- Last Active
- Roles
- Member
Comments
-
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
-
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…
-
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…
-
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?
-
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…
-
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
-
-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…
-
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…
-
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/…
-
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
-
Hi Jasmine, tensorflow example mnst是黑白的模型 所以"img_channel"要改"L"
-
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 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.
-
Try "pip install Pillow"
-
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.
-
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…
-
could you upload your littlenet models_720.nef
-
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…
-
The python examples is good start for learn APIs for config / inference / get result.
-
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…
-
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…
-
圖片是用 Netron 模型視覺化工具,可圖示了解模型結構
-
inpur_params.json 應修改"model_input_name": "inputs", https://www.kneron.com/forum/uploads/804/1CPO4FS3TT0J.jpg
-
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"])
-
Could you share your kl720_image_inference.py file?
-
ONNX is float , Kneron NPU is INT8 Kneron toolchain make model compression to int8 type (NEF). It have some lossy in inference.
-
If you use Vmware , try add CPU and RAM to setting https://www.kneron.com/forum/uploads/825/5DOEEFLHAHQQ.jpg
-
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…
-
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…
-
Same as the case cam_isi_post_host_yolov3_ssd_fd Just change the loop time number of cam_isi_pre_post_host_fcos.py Line 97: ret = user_test_single_yolo(device_index, 20)