Instructions to use altaidevorg/functiongemma-smarthome with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use altaidevorg/functiongemma-smarthome with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="altaidevorg/functiongemma-smarthome")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("altaidevorg/functiongemma-smarthome")
model = AutoModelForCausalLM.from_pretrained("altaidevorg/functiongemma-smarthome", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use altaidevorg/functiongemma-smarthome with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "altaidevorg/functiongemma-smarthome"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "altaidevorg/functiongemma-smarthome",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/altaidevorg/functiongemma-smarthome

SGLang

How to use altaidevorg/functiongemma-smarthome with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "altaidevorg/functiongemma-smarthome" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "altaidevorg/functiongemma-smarthome",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "altaidevorg/functiongemma-smarthome" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "altaidevorg/functiongemma-smarthome",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use altaidevorg/functiongemma-smarthome with Docker Model Runner:
```
docker model run hf.co/altaidevorg/functiongemma-smarthome
```

mys commited on Jan 28

Commit

af7768a

verified ·

1 Parent(s): 62219c1

Update README.md

Browse files

Files changed (1) hide show

README.md +8 -42

README.md CHANGED Viewed

@@ -9,50 +9,16 @@ tags:
 licence: license
 ---
-# Model Card for final_model_stable
-This model is a fine-tuned version of [google/functiongemma-270m-it](https://huggingface.co/google/functiongemma-270m-it).
-It has been trained using [TRL](https://github.com/huggingface/trl).
-## Quick start
-```python
-from transformers import pipeline
-question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
-generator = pipeline("text-generation", model="None", device="cuda")
-output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
-print(output["generated_text"])
-```
-## Training procedure
-This model was trained with SFT.
-### Framework versions
-- TRL: 0.23.0
-- Transformers: 4.56.2
-- Pytorch: 2.8.0
-- Datasets: 4.4.2
-- Tokenizers: 0.22.1
-## Citations
-Cite TRL as:
-```bibtex
-@misc{vonwerra2022trl,
-	title        = {{TRL: Transformer Reinforcement Learning}},
-	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
-	year         = 2020,
-	journal      = {GitHub repository},
-	publisher    = {GitHub},
-	howpublished = {\url{https://github.com/huggingface/trl}}
-}
-```

 licence: license
 ---
+# Model Card for functiongemma-smarthome
+[Dataset](https://huggingface.co/datasets/altaidevorg/smarthome-tool-calling-tiny) | [Notebook](https://github.com/altaidevorg/functiongemma-afterimage-demo/blob/main/fgemma-training.ipynb) | [Demo Video](https://www.youtube.com/watch?v=TJxtyrWSgo0)
+This model is a fine-tuned version of [google/functiongemma-270m-it](https://huggingface.co/google/functiongemma-270m-it) on a [custom tool-calling dataset](https://huggingface.co/datasets/altaidevorg/smarthome-tool-calling-tiny) synthetically generated with Afterimage, our purpose-built synthetic dataset generation engine.
+See the [demo video](https://www.youtube.com/watch?v=TJxtyrWSgo0).
+## What is Afterimage?
+Building custom Small Language Models (SLMs) starts with great data. Afterimage eliminates the tedious data preparation bottleneck by transforming your organization's unstructured documents into high-quality, LLM-ready Q&A sets, tool-calling datasets and/or other types  of structured datasets automatically. It is highly customizable and and aimed at transforming enterprises' way of customizing LLMs.
+## About ALTAI
+ALTAI is a secure, no-code platform that enables organizations to create, train, and deploy customized SLMs using their own internal documents. From "Letsearch" (RAG Engine) to on-premise deployment, we make LLM customization uncool again—simply effective. It can work 100% on-premise and requires 0 technical experience.