Default Avatar

Ethon Lin

Ethon Lin

About

Username
Ethon Lin
Joined
Visits
1,230
Last Active
Roles
Member, ModeratorsKL520, ModeratorsKL720, ModeratorsAIModel

Comments

  • 請參考文件的流程說明來進行量化(quantization)、編譯等流程 https://doc.kneron.com/docs/#toolchain/manual_1_overview/#12-workflow-overview
    in KL520 toolchain on docker Comment by Ethon Lin July 2024
  • Hello, Could you please provide a specific example to clarify where the issue is? For instance, provide an ONNX file and explain the steps you took, then describe the differences you found abnormal. These steps will help us assist you in analyzing t…
    in The loss of model conversion is too large Comment by Ethon Lin July 2024
  • Hello, Thanks for feedback. We did find some issues with the latest version 0.25.0 and are currently addressing them. Please hold on and wait for the next version.
    in Toolchain V0.25.0 Problem Comment by Ethon Lin July 2024
  • https://www.kneron.com/forum/discussion/comment/1828/#Comment_1828 如Maria所貼的規格，Toolchain最低會需要quad-core CPU，因為你使用的是 Intel Duo可能導致不相容，還請理解。
    in Docker環境下import ktc 發生core dump退出python Comment by Ethon Lin July 2024
  • Hello, It's recommended to build the DFUT tool by yourself instead of using the pre-built .sh file. Here is a reference link for building a console version of DFUT. https://www.kneron.com/forum/discussion/comment/1614/#Comment_1614
    in KL520 Update KDP2 firmware on Raspberry Pi4 Comment by Ethon Lin June 2024
  • Hello, 有關您的問題如下 是的，因為是耐能的NPU是INT8，建議運算的數值介於 -2^n ~ 2^n這個範圍才不會犧牲精確度，耐能本身都是採用-0.5~0.5這個範圍 因為投影片是依據過去修改的經驗整理出來的，所以行數可能與現在最新的版本有所差異。但最後一頁的概念就是要將訓練…
    in yolov7-tiny轉檔問題請教 Comment by Ethon Lin May 2024
  • Hello, 看訊息中有提到ArgMax，推測是模型中含有這個不支援的operator所造成 argmax是基於多個機率取最大值的運算，一般來說我們會放在後處理中再執行，所以要請你將onnx模型的最末端做調整，我們自己內部會是到Sigmoid的地方做結尾，剩餘的移到後處理來執行 https://w…
    in yolov7-tiny轉檔問題請教 Comment by Ethon Lin May 2024
  • Hi Nguyen, It's possible to burn the KL720 FW image to blank flash through UART and JTag (JLink), but you need to set the HW setup to change bootstrap to UART mode before burning by UART. The Jlink would be the easier way to access flash. To use USB…
    in KL720 Bootstrap config concern Comment by Ethon Lin May 2024
  • Hello, 若使用的是KL720的話，Mul與Sigmoid都要是可以轉換的才對，會方便請您提供onnx模型讓我們這邊分析後再跟您回報嗎?
    in yolov7 nef轉換 Comment by Ethon Lin May 2024
  • 您好，使用轉接頭將TypeC轉成TypeA是可行的，但從您的截圖來看，KL720依舊被降速成USB2.0 (High-Speed) 還請針對KL720的接線、轉接頭以及連接電腦時使用的usb port進行檢查，當中只要有一部分不符合usb3.0的規範整體就會被降速成USB2.0 若成功以USB3.0連接上的話，畫面…
    in Error: One or more selected dongless are not on USB Super-Speed. Comment by Ethon Lin April 2024
  • Hello, 相關資訊可能還是要請您向耐能業務或我們的代理商-全科聯絡，以便確認您的需求並回覆您的問題。
    in GPT的功能 是否有範例可以參考? Comment by Ethon Lin January 2024
  • Hello, 目前GPT相關功能只限商務客戶使用，若有相關需求請直接與耐能業務洽群。
    in GPT的功能 是否有範例可以參考? Comment by Ethon Lin January 2024
  • Hi Tri, The model "yolov5_pd/models_720.nef" is trained by Kneron. We provided the NEF file as an example for users, but we will not release the original ONNX model. If you want to train the yolov5/v7 model yourself, you can refer to the …
    in Load two models nef into KL720 Comment by Ethon Lin January 2024
  • Hello, Please refer to the following description for your question. It's recommended to combine different models into an NEF file before inference. You can then select the inference model by its ID. The issue with NEF combining occurred previously d…
    in Load two models nef into KL720 Comment by Ethon Lin January 2024
  • Hi Haresh, Maybe you can try to build a console DFUT instead of pre-built DFUT GUI tool we provided. step1: Download the latest PLUS here https://www.kneron.com/tw/support/developers/ Or download PLUS v2.2.0 directly by following link. https://www.…
    in I am not getting how to setup and use kl720 usb dongle in my pc. I am having linux 22.04 Comment by Ethon Lin January 2024
  • Hello, 請問是遇到插入KL520後執行DFUT卻找不到裝置的問題嗎? https://doc.kneron.com/docs/#plus_c/introduction/install_dependency/#213-using-zadig-to-install-driver 連結為Zadig的相關安裝說明，若要在Windows下操作dongle，請先透過Zadig安裝WinUSB，也提醒，…
    in KL-520 無法偵測~ Comment by Ethon Lin April 2023
  • Hi Alan, 推測可能燒錄過程中scpu有正常燒錄，所以系統相關的API (connection, update, etc..)可以正常執行，但是ncpu的部份燒錄失敗，因此在進行模型推論時才會發生錯誤 因為OS版本太多，我們只能挑幾個主流的版本來製作應用程式，所以DFUT目前並沒有規劃要支援額外的…
    in 在 Ubuntu 執行 KL720DemoGenericImageInferencePostYolo.py 失敗 Comment by Ethon Lin April 2023
  • Hello Alan, 因為過去曾多次發生用戶使用了不符合規格的線或品質不佳的線材導致一些不可預期的問題，所以強烈建議使用我們原裝驗證過的線材。 另外，我們發現你有在另外一篇討論中提到，同樣這隻dongle在Windows底下是可以正常透過DFUT燒錄的，想跟你確認一下目前的現…
    in 在 Ubuntu 執行 KL720DemoGenericImageInferencePostYolo.py 失敗 Comment by Ethon Lin April 2023
  • 記憶體配置的部分，每一包SDK在設定上都不太一樣，上面只是以mipi的API retrieve_new_buffer_for_mipi()來做說明，若以MIPI的應用來說，基本上應該不需要額外調整就可以使用。因為你有提到你們將相關流程改成DPI輸入，才會建議可以確認一下修改時是否有哪邊設定錯誤 從…
    in 有關kl520 2.0.1sdk的疑難排解 Comment by Ethon Lin March 2023
  • Hello, 有其他專案的窗口表示有收到你的反饋，將會另外與你聯絡，此篇討論將先close，後續若有其他問題也歡迎再提問
    in KL720 SD TF接口读写数据不成功 Comment by Ethon Lin March 2023
  • Hello, 請問使用的是KL720的哪一種裝置呢? dongle or demo board? 燒錄時選去的FW bin是哪一個版本呢? 若是用的是開發板，可以將UART接出後透過teraterm 或putty等終端機程式確認720的開機狀態是否有異常
    in KL720 更新 KDP2 Frimware 失敗後死機 Comment by Ethon Lin March 2023
  • Hello, Yolov5 & Yolov7我們都有持續在開發，但僅提供給商務合作相關的客戶使用，沒有在網路上免費公開，若有yolov7的使用需求，可以參考下面討論串的建議來修改模型 https://www.kneron.com/forum/discussion/comment/1421/#Comment_1421 Yolox的部份則是我們整…
    in 有關kl520 2.0.1sdk的疑難排解 Comment by Ethon Lin March 2023
  • Hello, 從描述來看推測是dpi傳圖片的部分可能沒有整合好，下面幾點想先和你確認 SDK v1.3是指kl520_sdk_v1.3.zip嗎? 在此版本上的整合都沒問題嗎? 除了版本新舊之外，是否有其他的考量導致想更新SDK呢? 因為SDK v1.3為舊架構SDK，與現在最新的 v2.1可以視為是完全不同…
    in 有關kl520 2.0.1sdk的疑難排解 Comment by Ethon Lin March 2023
  • Just left click mouse to select it. You can see the device you selected was highlight. https://www.kneron.com/forum/uploads/855/TMLOCP70HTX0.png
    in firmware update problem Comment by Ethon Lin March 2023
  • Hello, Please refer to our document. https://doc.kneron.com/docs/#plus_c/introduction/upgrade_ai_device_to_kdp2/#8-kl720-update-firmware-to-flash-memory
    in firmware update problem Comment by Ethon Lin March 2023
  • HI Wallace, We found an issue with the DFUT in version 2.1.1, which causes KL520 to fail to update from kdp to kdp2. Please download version 2.0.0 instead, and we will fix this error in next version. https://www.kneron.com/en/support/developers/ ht…
    in [KL520] Failure while updating FW to KDP2 Comment by Ethon Lin March 2023
  • Hello, Kneron's quantization is used for floating to INT8 format, you can get more detail from following link. https://doc.kneron.com/docs/#toolchain/manual_4_bie/#41-quantization It's possible to get extra loss if you are using quantizing model to…
    in kl720 compile and quantization issuse Comment by Ethon Lin February 2023
  • Hello, 能夠轉出NEF即代表成功將模型轉換成KL520能支援的格式，但執行推論時須配合模型訓練、轉換過程中的相關設定，因此需請你確認相關參數來比對答案與原始模型是否有差異 USB BOOT與FLASH BOOT的差異在於FW的取得來源，FLASH BOOT的情況下需要先將FW燒錄至flash後…
    in 從0開始學習KL520 AI訓練 Comment by Ethon Lin February 2023
  • Hello, 下面幾個檢查方向可供你參考 pytorch的模型轉成onnx後還需要再跑一次converter，來將模型轉換成符合耐能NPU的格式，詳細可參考耐能文件的3.1.2. Pytorch to ONNX https://doc.kneron.com/docs/#toolchain/manual_3_onnx/#31-model-conversion-and-optim…
    in Yolov7-tiny轉檔問題 Comment by Ethon Lin February 2023
  • 已於郵件另行回覆，若仍有相關疑問，歡迎另開新的討論貼文或直接向商務/技術窗口提出。
    in SDK M5S 2.4.5中获取模型推理结果的问题 Comment by Ethon Lin February 2023

© Kneron, Inc.