Awesome

Home LLM

This project provides the required "glue" components to control your Home Assistant installation with a completely local Large Language Model acting as a personal assistant. The goal is to provide a drop in solution to be used as a "conversation agent" component by Home Assistant. The 2 main pieces of this solution are the Home LLM model and Local LLM Conversation integration.

Quick Start

Please see the Setup Guide for more information on installation.

Local LLM Conversation Integration

The latest version of this integration requires Home Assistant 2024.12.3 or newer

In order to integrate with Home Assistant, we provide a custom component that exposes the locally running LLM as a "conversation agent".

This component can be interacted with in a few ways:

• using a chat interface so you can chat with it.
• integrating with Speech-to-Text and Text-to-Speech addons so you can just speak to it.

The integration can either run the model in 2 different ways:

1. Directly as part of the Home Assistant software using llama-cpp-python
2. On a separate machine using one of the following backends:
   • Ollama (easier)
   • LocalAI via the Generic OpenAI backend (easier)
   • oobabooga/text-generation-webui project (advanced)
   • llama.cpp example server (advanced)

Home LLM Model

The "Home" models are a fine tuning of various Large Languages Models that are under 5B parameters. The models are able to control devices in the user's house as well as perform basic question and answering. The fine tuning dataset is a custom synthetic dataset designed to teach the model function calling based on the device information in the context.

The latest models can be found on HuggingFace:
3B v3 (Based on StableLM-Zephyr-3B): https://huggingface.co/acon96/Home-3B-v3-GGUF (Zephyr prompt format)
1B v3 (Based on TinyLlama-1.1B): https://huggingface.co/acon96/Home-1B-v3-GGUF (Zephyr prompt format)

Non English experiments:
German, French, & Spanish (3B): https://huggingface.co/acon96/stablehome-multilingual-experimental
Polish (1B): https://huggingface.co/acon96/tinyhome-polish-experimental

3B v2 (Based on Phi-2): https://huggingface.co/acon96/Home-3B-v2-GGUF (ChatML prompt format)
1B v2 (Based on Phi-1.5): https://huggingface.co/acon96/Home-1B-v2-GGUF (ChatML prompt format)
1B v1 (Based on Phi-1.5): https://huggingface.co/acon96/Home-1B-v1-GGUF (ChatML prompt format)

NOTE: The models below are only compatible with version 0.2.17 and older! 3B v1 (Based on Phi-2): https://huggingface.co/acon96/Home-3B-v1-GGUF (ChatML prompt format)

The model is quantized using Llama.cpp in order to enable running the model in super low resource environments that are common with Home Assistant installations such as Raspberry Pis.

The model can be used as an "instruct" type model using the ChatML or Zephyr prompt format (depends on the model). The system prompt is used to provide information about the state of the Home Assistant installation including available devices and callable services.

Example "system" prompt:

You are 'Al', a helpful AI Assistant that controls the devices in a house. Complete the following task as instructed with the information provided only.
The current time and date is 08:12 AM on Thursday March 14, 2024
Services: light.turn_off(), light.turn_on(brightness,rgb_color), fan.turn_on(), fan.turn_off()
Devices:
light.office 'Office Light' = on;80%
fan.office 'Office fan' = off
light.kitchen 'Kitchen Light' = on;80%;red
light.bedroom 'Bedroom Light' = off

For more about how the model is prompted see Model Prompting

Output from the model will consist of a response that should be relayed back to the user, along with an optional code block that will invoke different Home Assistant "services". The output format from the model for function calling is as follows:

turning on the kitchen lights for you now
```homeassistant
{ "service": "light.turn_on", "target_device": "light.kitchen" }
```

Due to the mix of data used during fine tuning, the 3B model is also capable of basic instruct and QA tasks. For example, the model is able to perform basic logic tasks such as the following (Zephyr prompt format shown):

<|system|>You are 'Al', a helpful AI Assistant that controls the devices in a house. Complete the following task as instructed with the information provided only.
*snip*
<|endoftext|>
<|user|>
if mary is 7 years old, and I am 3 years older than her. how old am I?<|endoftext|>
<|assistant|>
If Mary is 7 years old, then you are 10 years old (7+3=10).<|endoftext|>

Synthetic Dataset

The synthetic dataset is aimed at covering basic day to day operations in home assistant such as turning devices on and off. The supported entity types are: light, fan, cover, lock, media_player, climate, switch

The dataset is available on HuggingFace: https://huggingface.co/datasets/acon96/Home-Assistant-Requests
The source for the dataset is in the data of this repository.

Training

If you want to prepare your own training environment, see the details on how to do it in the Training Guide document.

The 3B model was trained as a full fine-tuning on 2x RTX 4090 (48GB). Training time took approximately 28 hours. It was trained on the --large dataset variant.

accelerate launch --config_file fsdp_config.yaml train.py \
    --run_name home-3b \
    --base_model stabilityai/stablelm-zephyr-3b \
    --bf16 \
    --train_dataset data/home_assistant_train.jsonl \
    --learning_rate 1e-5 \
    --batch_size 64 \
    --epochs 1 \
    --micro_batch_size 2 \
    --gradient_checkpointing \
    --group_by_length \
    --ctx_size 2048 \
    --save_steps 50 \
    --save_total_limit 10 \
    --eval_steps 100 \
    --logging_steps 2

The 1B model was trained as a full fine-tuning on an RTX 3090 (24GB). Training took approximately 2 hours. It was trained on the --medium dataset variant.

python3 train.py \
    --run_name home-1b \
    --base_model TinyLlama/TinyLlama-1.1B-Chat-v1.0 \
    --bf16 \
    --train_dataset data/home_assistant_train.jsonl \
    --test_dataset data/home_assistant_test.jsonl \
    --learning_rate 2e-5 \
    --batch_size 32 \
    --micro_batch_size 8 \
    --gradient_checkpointing \
    --group_by_length \
    --ctx_size 2048 \
    --save_steps 100 \
    --save_total_limit 10
    --prefix_ids 29966,29989,465,22137,29989,29958,13 \
    --suffix_ids 2

Home Assistant Addon

In order to facilitate running the project entirely on the system where Home Assistant is installed, there is an experimental Home Assistant Add-on that runs the oobabooga/text-generation-webui to connect to using the "remote" backend options. The addon can be found in the addon/ directory.

Version History