Instructions to use shibatch/tinygemma3-2m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use shibatch/tinygemma3-2m with Transformers:

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

pipe = pipeline("text-generation", model="shibatch/tinygemma3-2m")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("shibatch/tinygemma3-2m", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use shibatch/tinygemma3-2m with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf shibatch/tinygemma3-2m:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf shibatch/tinygemma3-2m:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf shibatch/tinygemma3-2m:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf shibatch/tinygemma3-2m:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf shibatch/tinygemma3-2m:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf shibatch/tinygemma3-2m:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf shibatch/tinygemma3-2m:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf shibatch/tinygemma3-2m:Q4_K_M

Use Docker

docker model run hf.co/shibatch/tinygemma3-2m:Q4_K_M

LM Studio
Jan

vLLM

How to use shibatch/tinygemma3-2m with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "shibatch/tinygemma3-2m"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "shibatch/tinygemma3-2m",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/shibatch/tinygemma3-2m:Q4_K_M

SGLang

How to use shibatch/tinygemma3-2m 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 "shibatch/tinygemma3-2m" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "shibatch/tinygemma3-2m",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "shibatch/tinygemma3-2m" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "shibatch/tinygemma3-2m",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use shibatch/tinygemma3-2m with Ollama:
```
ollama run hf.co/shibatch/tinygemma3-2m:Q4_K_M
```
Unsloth Desktop
Docker Model Runner
How to use shibatch/tinygemma3-2m with Docker Model Runner:
```
docker model run hf.co/shibatch/tinygemma3-2m:Q4_K_M
```

Lemonade

How to use shibatch/tinygemma3-2m with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull shibatch/tinygemma3-2m:Q4_K_M

Run and chat with the model

lemonade run user.tinygemma3-2m-Q4_K_M

List all available models

lemonade list

Atomic Chat

tinygemma3-2m

File size: 896 Bytes

554dc63

from pathlib import Path

import torch
from transformers import Gemma3ForCausalLM, PreTrainedTokenizerFast


MODEL_DIR = Path(__file__).resolve().parent / "hf"
PROMPT = "Once upon"


def main() -> None:
    tokenizer = PreTrainedTokenizerFast.from_pretrained(MODEL_DIR)
    model = Gemma3ForCausalLM.from_pretrained(
        MODEL_DIR,
        dtype=torch.float32,
    ).eval()

    input_ids = torch.tensor(
        [
            [tokenizer.bos_token_id]
            + tokenizer.encode(PROMPT, add_special_tokens=False)
        ]
    )

    with torch.no_grad():
        output = model.generate(
            input_ids,
            max_new_tokens=100,
            do_sample=False,
            pad_token_id=tokenizer.pad_token_id,
            eos_token_id=tokenizer.eos_token_id,
        )

    print(tokenizer.decode(output[0], skip_special_tokens=True))


if __name__ == "__main__":
    main()