# ![huggy-face](data:image/png;base64,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)Hugging Face Optimum + ONNX-RT

本節示範如何在 Snapdragon X 平台上使用 [Hugging Face Optimum 和 ONNX Runtime](https://huggingface.co/docs/optimum/v1.2.1/en/onnxruntime/modeling_ort)。

## 設置

以系統管理員模式開啟 PowerShell，並執行以下命令以安裝相依性。為了教學，必須進行此安裝，大約需要 5 分鐘。若需要與安裝有關的詳細資訊，請參閱 [ort_setup.ps1](https://raw.githubusercontent.com/quic/wos-ai/refs/heads/main/Scripts/ort_setup.ps1)。

1. 設定 $DIR\_PATH（預設為 `C:\WoS_AI`），使用者可以變更為需要的路徑。

$DIR_PATH  = "C:\WoS_AI"
        Copy to clipboard
2. 執行下列命令，下載安裝指令碼。若已下載，則可跳過此步驟。

if (!(Test-Path $DIR_PATH\Downloads\Setup_Scripts)) {mkdir $DIR_PATH\Downloads\Setup_Scripts}
        Invoke-WebRequest -O ort_setup.ps1 https://raw.githubusercontent.com/quic/wos-ai/refs/heads/main/Scripts/ort_setup.ps1
        Move-Item -Path ".\ort_setup.ps1" -Destination $DIR_PATH\Downloads\Setup_Scripts -Force
        Copy to clipboard
3. 執行以下指令來安裝 Hugging Face Optimum 的必要條件。

cd $DIR_PATH
        powershell -command "&{. .\Downloads\Setup_Scripts\ort_setup.ps1; ORT_HF_Setup -rootDirPath $DIR_PATH}"
        Copy to clipboard

    以下相依性已下載並安裝：

> 
> 
> - Python amd64 3.12.6 版本
>     - Visual Studio 可轉散發套件版本 14.44.35208
>     - 創建一個虛擬環境（SDX\_ORT\_HF\_ENV），包含 Hugging Face 所需的必要庫。

備註

如果在系統上無法執行指令碼，請依據下列所示執行以下命令，並輸入「A」（全部 Yes）。

Set-ExecutionPolicy RemoteSigned
        Copy to clipboard
4. 設置完成後，$DIR\_PATH 中將有以下資料夾結構：

![../../_images/folders_structure_ort.png](data:image/png;base64,UklGRr4GAABXRUJQVlA4TLIGAAAv2YAeAP/BIJIkJ/1qIkE1PrCQswgcYINNZNtOVswnR8/UlPRYyFkEDmDQSJKiEvESmPZFvbf3w+zl5j/w/xMbXjgI8FNALrQ9NYwIZlj1UKOE0EECKxYuZgwSDvwkDwEFr50NL1eAlQHCqmetmOHRgAo5NRH8Lo/VjpU5mKBHZfbQoEY3m0AYx8n3ZfQdULS2PW7USDYg28OaEjfqIhXE4lKxg0JqU93/bVnf/8vxTDTr6YnoPwRJcuM2oxTjU1lQkAQQyAfkvP2UR24rSrst3ZpdvErhhNhXvRAhp2V87TlMm8QRAbPE4+iTslEvtTHmbU5bXG3ON/XamEe16mLmUNkgpep8Rt1AZ2Orjz9EEYxZB2k9fP+ios60/ff9/D0pLfEDF8DY3FyYYUVAHeNhxB7qaFOdRapDJXuyWyume7ukQXXJi6hDfRqIX4Wks0P1ph31rNc3u9m8WYxfgvt4eDDj9hV5z2V8kxi91YtodmhoFjubYKcaBFB7lg6ME7hIDAGLm2eaeFEfkoDADeaxps7Yp1kE45a5Z7P7sKmf//Gobt5UOGNEjuctT579frVBoIs8r3F+GXvWgIqgfiVCOpbZJy81KmbnoE8z75dcat4dKvd5XzGgdDPkPWKq13vS7OKapEnLwVq9FuZHteyK0j/x0rSB2GQG9DlpY8ZwULIa4MTMbDH2AOmc4MHvLkD2kVxCe49+ngAPNgPK4+fAfdj5RTMwaqk6OIvQIZrNK3Bk3q+nOqkK6+wIJJvWyx5uO4w9cM3J39WIGVYIEDk7JSOW2t4L9gIYqdrk99y9SnX76voVOdlL6jfecv3KB0CQbV9dUzKX21TskTGRaUMxOJ+c2wxHiirGsDTnvoxg8gWanzsx6hQsByu+74TwBGIFxNJMNg4h8/THgALG17E4YMZC83N1pk0LizbznlLizaG6CAsEsvlPoqLew8uoXnWnhli2TfUCmCx2F+3XQT2scRq3LDZKml5iIqTeElGoJJTqCUU6v/1t82JfHQPmrFsW3HpfIlzk/pz9dqic8DIHt6yw1fAtO2pQhesDfQT1JdpRkJLJz1XnYR+WrP9G1YE2T1ITL/Ei9ytd2PvUJn3pyWUW1d0EcUczl3ciFh7VLYjHrCc/K2BasgxcoIEJr51fzDqi1aFgwOcgh+/F/kFHq1DuA83le1pz+tSKUeewfv4ck/dbIaYXqfGhHFifcnSQyQ4iQsbe7DxR5KY6FT9z51BSyeX9LshkQhlyedWRM9QGKCVln2x1qp4zpVXw+dyADKAR33OhHE+IxGc6qVwfKTPMr8y/S4M0g9tXBDvVFK4xYTlbzWIef9lXBdkR0daPGs8csiwoPLTH4E0lyCTCzzqnOrujUTN5v0zqYoEiBBcurdhHnenLCKZzorCR+IW+g8W7X7n8SaTwBayPr5nCZ++5hauPrwtJ30sIzhAEPwHoSSQeofcbovXx926qF0B62UnU8GGHlXUxvdYFodnUzpNY3AlPGsppRfVMY0fqAmjPOZjX7NKcuwit40Wpj9/bUUHO424oqMz8WID1TGc2pIXqYU07Icvy5P6mzrz8UudQhxVoNjVziytMfTyQw+1Gja5hA/R2ow/0+0XgpJuBLCpNNenu2OqS1MdRAS4IfuzwNIFW9B52L8S7ecim8MKTnnPi/N83NTnrfFljH3WoVv1/obJigzwSvLiVF9UdH+rg7bMudyTuBNnCHb/6uGlDtue8wJiWxG/xyG5rM5R4sE+8lOapl5cRKcTevvaS9JwzfeS08EXggqBf6mLiXVRvrZe055z2kadECHRTSFZkp3U46VyewwwuztDObNptTXOkwGLaIpZHOFIyu5Bh4cvqTJcT/B6zFpHtI7fckJJ6yP4KiUtnBCTttOec6yNvdrhQS3vOiYAo79RtLpSYn6tx7R7Mo/pSGazweUx7zmXlXohwioPqRg1GezbIPZ+Cg/XHWe7CKQ7mSUp/nw7y4qIOFZTcT/Bl8/emlu7TgFnmqCFNZopaU53BtEka35/uuHieb0bP96fb0xyGVFX9c5Cqi5gXWhHoPpsNB1NJPsP6JK1rQsf3goHpT2cKSBdxEMu2UXOow1/eDNKSFQtyXK1fgiW1Sq7XuSdWLIDtTz+9tQ+OdOAzeCDm4WaNY6sXv8z1p5Ma9crDhQG/x3WYWxHB4RdA6uOOG5J7g5xExie4/nTVefjICa5jkLK52FcWPdyJf/ASRlycpP1wL1Rdvj8dxEaul53fehGZU2e/9teuDds1ZjAts4qcH65HL/9/9CAB)
    - Downloads：完成安裝此資料夾儲存需要的所有檔案，例如 Python 安裝程式、安裝指令碼等。
    - Python\_Env：此目錄包含為 EP 建立的虛擬環境 (SDX\_ORT\_HF\_ENV)。
    - Debug\_Logs：此目錄包含與安裝對應的紀錄。
    - Models：此目錄包含模型子目錄及其對應的元件（例如 Mobilenet\_v2）。

## 教學

**CPU EP — 範例應用程式工作流程**

> 
> 
> 1. 開啟 PowerShell 並執行以下命令以設定 $DIR\_PATH 並啟動 Python 虛擬環境。確保 $DIR\_PATH 與安裝時使用的路徑相同
> 
> 
> ${DIR_PATH} = "C:\WoS_AI"
>         Copy to clipboard
> 
> 
> powershell -NoExit -command "&{cd $DIR_PATH; . .\Downloads\Setup_Scripts\ort_setup.ps1; Activate_ORT_HF_VENV -rootDirPath $DIR_PATH}"
>         Copy to clipboard
> 2. 在工作目錄中執行以下命令來建立 hf\_optimum\_cpu.py 檔案。
> 
> 
> notepad.exe hf_optimum_cpu.py
>         Copy to clipboard
> 3. 將以下程式碼複製到 *hf\_optimum\_cpu.py* 檔案中並儲存。
> 
> 
> # File_name : hf_optimum_cpu.py
>         
>         # Import the ORT Model for Sequence classification
>         from transformers import AutoTokenizer, pipeline
>         from optimum.onnxruntime import ORTModelForSequenceClassification
>         
>         # Load the Pre trained model and update the tokenizer
>         model = ORTModelForSequenceClassification.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english", from_transformers=True)
>         tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")
>         
>         # setup the pipeline and classifier inputs
>         onnx_classifier = pipeline("text-classification", model=model, tokenizer=tokenizer)
>         
>         # Run the inference
>         result = onnx_classifier("This is a great model")
>         print("Predicted Result : ", result)
>         Copy to clipboard
> 4. 在工作目錄中執行 *hf\_optimum\_cpu.py* Python 腳本。
> 
> 
> Python .\hf_optimum_cpu.py
>         Copy to clipboard
> 
> 
>     此指令碼是使用 Python 常用工具庫載入預訓練模型，並將其傳送至 HF\_Optimum\_CPU EP 進行推論。之後，處理 HF\_Optimum\_CPU EP 推論結果，將指定文字的分類顯示為正向，如下圖所示。
> 
> ![../../_images/hf-optimum-cpu-py-screenshot.png](data:image/png;base64,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)

## 下一步

我們已成功執行 HuggingFace 的 DistilBert 模型，並使用 ORT CPU EP 將指定文字進行分類。現在，您可以進入下一節，其中介紹了可以針對 AI 工作負載獲得最佳每瓦效能的 Qualcomm AI Runtime SDK。

## 常見問題

> 
> 
> - **支援哪一個 Python 版本？**
> 
>     - Python arm64 變體目前不支援 ONNX Runtime for HF EP。
>     - ONNX Runtime HF EP 需要 Python amd64 版本 3.8.10 至 3.12.x
>     - 已使用 [python-3.12.6-amd64](https://www.python.org/ftp/python/3.12.6/python-3.12.6-amd64.exe) 驗證教學範例

Last Published: Dec 16, 2025

[Previous Topic
QNN EP](https://docs.qualcomm.com/bundle/publicresource/80-62010-1TC/topics/ort-qnn-ep.md) [Next Topic
Qualcomm AI Runtime (QNN)](https://docs.qualcomm.com/bundle/publicresource/80-62010-1TC/topics/qnn.md)