Skip to content

Latest commit

 

History

History
82 lines (60 loc) · 4.01 KB

INSTALL_win_arc.md

File metadata and controls

82 lines (60 loc) · 4.01 KB
Raw

Setup Guide for Windows with Intel Arc A-Series GPU

  1. Installation
  2. Start the Service

Installation

Download Langchain-Chatchat

Download the Langchain-Chatchat with IPEX-LLM integrations from this link. Unzip the content into a directory, e.g. C:\Users\arda\Downloads\Langchain-Chatchat-ipex-llm.

Install Prerequisites

Visit the Install IPEX-LLM on Windows with Intel GPU Guide, and follow Install Prerequisites to update GPU driver (optional), and install Conda.

Install Python Dependencies

1. Create a Conda Environment

Open Miniforge Prompt, and run the following commands to create a new python environment:

conda create -n ipex-llm-langchain-chatchat python=3.11 libuv 
conda activate ipex-llm-langchain-chatchat

2. Install ipex-llm

pip install --pre --upgrade ipex-llm[xpu] --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/
pip install --pre --upgrade torchaudio==2.1.0a0  --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/

Note

You can also use --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/cn/.

3. Install Langchain-Chatchat Dependencies

Switch to the root directory of Langchain-Chatchat you've downloaded (refer to the download section), and install the dependencies with the commands below. Note: In the example commands we assume the root directory is C:\Users\arda\Downloads\Langchain-Chatchat-ipex-llm. Remember to change it to your own path.

cd C:\Users\arda\Downloads\Langchain-Chatchat-ipex-llm
pip install -r requirements_ipex_llm.txt 
pip install -r requirements_api_ipex_llm.txt
pip install -r requirements_webui.txt

Configuration

  • In root directory of Langchain-Chatchat, run the following command to create a config: 
    python copy_config_example.py
  • Edit the file configs/model_config.py, change MODEL_ROOT_PATH to the absolute path of the parent directory where all the downloaded models (LLMs, embedding models, ranking models, etc.) are stored.

Download Models

Download the models and place them in the directory MODEL_ROOT_PATH (refer to details in Configuration section).

Currently, we support only the LLM/embedding models specified in the table below. You can download these models using the link provided in the table. Note: Ensure the model folder name matches the last segment of the model ID following "/", for example, for THUDM/chatglm3-6b, the model folder name should be chatglm3-6b.

Model Category download link
THUDM/chatglm3-6b Chinese LLM HF or ModelScope
meta-llama/Llama-2-7b-chat-hf English LLM HF
BAAI/bge-large-zh-v1.5 Chinese Embedding HF
BAAI/bge-large-en-v1.5 English Embedding HF

Start the Service

In Miniforge Prompt, under the root directory of Langchain-Chatchat, run the following commands:

conda activate ipex-llm-langchain-chatchat

set SYCL_CACHE_PERSISTENT=1
set BIGDL_QUANTIZE_KV_CACHE=1
set BIGDL_LLM_XMX_DISABLED=1

set no_proxy=localhost,127.0.0.1
set BIGDL_IMPORT_IPEX=0

python startup.py -a

Note

The above configurations lead to optimal performance for Intel Arc™ A-Series Graphics with the exception of Intel Arc™ A300-Series or Pro A60.

You can find the Web UI's URL printed on the terminal logs, e.g. http://localhost:8501/.

Open a browser and navigate to the URL to use the Web UI.