Lab 0 - Installation

When we enter the Lab, we need to configure the relevant environment :

1. Python 3.11+

It is recommended to use miniforge to configure your Python environment

To configure miniforge, please refer to https://github.com/conda-forge/miniforge

After configuring miniforge, run the following command in Power Shell

conda create -n pyenv python==3.11.8 -y

conda activate pyenv

2. Install Prompt flow SDK

In Lab 1, we use Prompt flow, so you need to configure the Prompt flow SDK.

pip install promptflow --upgrade

You can check promptflow sdk in this command

pf --version

3. Install Visual Studio Code Prompt flow Extension

4. Intel NPU Acceleration Library

Intel's new generation processors support NPU. If you want to use NPU to run LLMs / SLMs locally, you can use Intel NPU Acceleration Library . If you want to know more, you can read https://github.com/microsoft/PhiCookBook/blob/main/md/01.Introduction/03/AIPC_Inference.md.

Install Intel NPU Acceleration Library in bash

pip install intel-npu-acceleration-library

Note: Please note that this library supports transformers 4.40.2 , please confirm the version

5. Other Python Library

create requirements.txt and add this content

notebook
numpy 
scipy 
scikit-learn 
matplotlib 
pandas 
pillow 
graphviz

6. Install NVM

install nvm in Powershell

winget install -e --id CoreyButler.NVMforWindows

install nodejs 18.20

nvm install 18.20.0

nvm use 18.20.0

7. Install Visual Studio Code Development Support

npm install --global yo generator-code

Congratulations! You have successfully configured the SDK. Next, proceed to the hands-on steps.