Skip to content

Latest commit

 

History

History
125 lines (105 loc) · 4.11 KB

installation.md

File metadata and controls

125 lines (105 loc) · 4.11 KB

Installing MemoryScope

I. Install with docker [Recommended] [x86_64]

  1. Clone the repository and edit settings

    # clone project
git clone https://github.com/modelscope/memoryscope
cd memoryscope
# edit configuration, e.g. add api keys
vim memoryscope/core/config/demo_config.yaml

  2. Build Docker image

    sudo docker build --network=host -t memoryscope .

  3. Launch Docker container (ARM architecture not yet supported)

    sudo docker run -it --rm --net=host memoryscope

Important

To inspect memory shift during the conversation, modify command in step 3 to sudo docker run -it --name=memoryscope_container --rm --net=host memoryscope;
Then start a new terminal window and execute sudo docker exec -it memoryscope_container python quick-start-demo.py --config_path=memoryscope/core/config/demo_config_zh.yaml;
In the second window, input /list_memory refresh_time=5 to inspect memory

II. Install with docker compose [Recommended] [x86_64]

  1. Clone the repository and edit settings

    # clone project
git clone https://github.com/modelscope/memoryscope
cd memoryscope
# edit configuration, e.g. add api keys
vim memoryscope/core/config/demo_config.yaml

  2. Edit docker-compose.yml to change environment variable.

    OPENAI_API_KEY: "sk-0000000000"

  3. Run docker-compose run memory_scope_main to build and launch the memory-scope cli interface. (ARM architecture not yet supported)

III. Install from PyPI

  1. Install from PyPI

    pip install memoryscope

  2. Run Elasticsearch service, refer to elasticsearch documents. The docker method is recommended:

    sudo docker run -p 9200:9200 \
-e "discovery.type=single-node" \
-e "xpack.security.enabled=false" \
-e "xpack.license.self_generated.type=trial" \
docker.elastic.co/elasticsearch/elasticsearch:8.13.2

  3. Test Chinese / Dashscope Configuration

    export DASHSCOPE_API_KEY="sk-0000000000"
memoryscope --language="cn" \
        --memory_chat_class="cli_memory_chat" \
        --human_name="用户" \
        --assistant_name="AI" \
        --generation_backend="dashscope_generation" \
        --generation_model="qwen-max" \
        --embedding_backend="dashscope_embedding" \
        --embedding_model="text-embedding-v2" \
        --enable_ranker=True \
        --rank_backend="dashscope_rank" \
        --rank_model="gte-rerank"

  4. Test English / OpenAI Configuration

    export OPENAI_API_KEY="sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
memoryscope --language="en" \
        --memory_chat_class="cli_memory_chat" \
        --human_name="User" \
        --assistant_name="AI" \
        --generation_backend="openai_generation" \
        --generation_model="gpt-4o" \
        --embedding_backend="openai_embedding" \
        --embedding_model="text-embedding-3-small" \
        --enable_ranker=False

IV. Install from source

  1. Clone the repository and edit settings

    # clone project
git clone https://github.com/modelscope/memoryscope
cd memoryscope
# edit configuration, e.g. add api keys
vim memoryscope/core/config/demo_config.yaml

  2. Install

    pip install -e .

  3. Run Elasticsearch service, refer to elasticsearch documents. The docker method is recommended:

    sudo docker run -p 9200:9200 \
-e "discovery.type=single-node" \
-e "xpack.security.enabled=false" \
-e "xpack.license.self_generated.type=trial" \
docker.elastic.co/elasticsearch/elasticsearch:8.13.2

  4. Launch memoryscope, also refer to cli documents

    export OPENAI_API_KEY="sk-0000000000"
python quick-start-demo.py --config_path=memoryscope/core/config/demo_config_zh.yaml