Instructions to use CuriousDragon/functiongemma-270m-tiny-agent with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use CuriousDragon/functiongemma-270m-tiny-agent with Transformers:

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

pipe = pipeline("text-generation", model="CuriousDragon/functiongemma-270m-tiny-agent")
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("CuriousDragon/functiongemma-270m-tiny-agent")
model = AutoModelForCausalLM.from_pretrained("CuriousDragon/functiongemma-270m-tiny-agent", 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 CuriousDragon/functiongemma-270m-tiny-agent with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "CuriousDragon/functiongemma-270m-tiny-agent"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CuriousDragon/functiongemma-270m-tiny-agent",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/CuriousDragon/functiongemma-270m-tiny-agent

SGLang

How to use CuriousDragon/functiongemma-270m-tiny-agent 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 "CuriousDragon/functiongemma-270m-tiny-agent" \
    --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": "CuriousDragon/functiongemma-270m-tiny-agent",
		"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 "CuriousDragon/functiongemma-270m-tiny-agent" \
        --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": "CuriousDragon/functiongemma-270m-tiny-agent",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use CuriousDragon/functiongemma-270m-tiny-agent with Docker Model Runner:
```
docker model run hf.co/CuriousDragon/functiongemma-270m-tiny-agent
```

Tiny Agent: FunctionGemma-270m-IT (Fine-Tuned)

This is a fine-tuned version of google/functiongemma-270m-it optimized for reliable function calling. It was trained as part of the "Tiny Agent Lab" project to distill the capabilities of larger models into a highly efficient 270M parameter model.

Model Description

Model Type: Causal LM (Gemma)
Language(s): English
License: Gemma Terms of Use
Finetuned from: google/functiongemma-270m-it

Capabilities

This model is designed to:

Detect User Intent: Accurately identify when a tool call is needed.
Generate Function Calls: Output valid <start_function_call> XML/JSON blocks.
Refuse Out-of-Scope Requests: Politely decline requests for which no tool is available.
Ask Clarification: Request missing parameter values interactively.

Performance (V4 Evaluation)

On a held-out test set of 100 diverse queries:

Overall Accuracy: 71%
Tool Selection Precision: 88%
Tool Selection Recall: 94%
F1 Score: 0.91

Usage

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_id = "CuriousDragon/functiongemma-270m-tiny-agent"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", torch_dtype=torch.float16)

# ... (Add your inference code here)

Intended Use

This model is intended for research and educational purposes in building efficient agentic systems. It works best when provided with a clear system prompt defining the available tools.

Downloads last month: 18

Safetensors

Model size

0.3B params

Tensor type

F16

Model tree for CuriousDragon/functiongemma-270m-tiny-agent

Base model

google/functiongemma-270m-it

Finetuned

(472)

this model

Finetunes

1 model