llama-index

llama-index

CONTAINERS IMAGES RUN BUILD

• llama-index from https://www.llamaindex.ai/

Starting llamaindex container

jetson-containers run $(autotag llama-index)

This will start the ollama server as well as Jupyter Lab server inside the container.

Running a RAG example with Ollama

This is based on the official tutorial for local models.

Jupyter Notebook Version

When you run start the llama-index container, you should see lines like this on the terminal.

JupyterLab URL:   http://192.168.1.10:8888 (password "nvidia")
JupyterLab logs:  /data/logs/jupyter.log

On your Jetson desktop GUI, or on a PC on the same network as Jetson, open your web browser and access the address. When prompted, type the password nvidia and log in.

Jupyter Lab UI should show up, with LlamaIndex_Local-Models.ipynb listed in the left navigator pane - open it, and follow the guide in the Jupyter notebook.

Python Version

After starting the llamaindex container, you should be on root@<hostname> console. First, download the Llama2 model using ollama

ollama pull llama2

This downloads the default 7-billion parameter Llama2 model - you can optionally specify ollma2:13b and ollma2:70b for other variations, and change the Python script (line 13) accordingly. Then type the following to start the sample Python script:

python3 /opt/llama-index/llamaindex_starter.py

CONTAINERS

llama-index
Aliases	llama-index:main
Requires	L4T ['>=34.1.0']
Dependencies	build-essential cuda:12.2 cudnn python numpy cmake onnx pytorch
Dependants	llama-index:samples
Dockerfile	Dockerfile
Images	dustynv/llama-index:r35.4.1 (2024-05-23, 5.5GB) dustynv/llama-index:r36.2.0 (2024-04-30, 6.2GB) dustynv/llama-index:r36.3.0 (2024-05-23, 5.5GB)

CONTAINER IMAGES

Repository/Tag	Date	Arch	Size
dustynv/llama-index:r35.4.1	2024-05-23	arm64	5.5GB
dustynv/llama-index:r36.2.0	2024-04-30	arm64	6.2GB
dustynv/llama-index:r36.3.0	2024-05-23	arm64	5.5GB

_{Container images are compatible with other minor versions of JetPack/L4T:
    • L4T R32.7 containers can run on other versions of L4T R32.7 (JetPack 4.6+)
    • L4T R35.x containers can run on other versions of L4T R35.x (JetPack 5.1+)}

RUN CONTAINER

To start the container, you can use jetson-containers run and autotag, or manually put together a docker run command:

# automatically pull or build a compatible container image
jetson-containers run $(autotag llama-index)

# or explicitly specify one of the container images above
jetson-containers run dustynv/llama-index:r35.4.1

# or if using 'docker run' (specify image and mounts/ect)
sudo docker run --runtime nvidia -it --rm --network=host dustynv/llama-index:r35.4.1

^{jetson-containers run forwards arguments to docker run with some defaults added (like --runtime nvidia, mounts a /data cache, and detects devices)
autotag finds a container image that's compatible with your version of JetPack/L4T - either locally, pulled from a registry, or by building it.}

To mount your own directories into the container, use the -v or --volume flags:

jetson-containers run -v /path/on/host:/path/in/container $(autotag llama-index)

To launch the container running a command, as opposed to an interactive shell:

jetson-containers run $(autotag llama-index) my_app --abc xyz

You can pass any options to it that you would to docker run, and it'll print out the full command that it constructs before executing it.

BUILD CONTAINER

If you use autotag as shown above, it'll ask to build the container for you if needed. To manually build it, first do the system setup, then run:

jetson-containers build llama-index

The dependencies from above will be built into the container, and it'll be tested during. Run it with --help for build options.