Instructions to use google/gemma-3-270m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use google/gemma-3-270m with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="google/gemma-3-270m")# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-3-270m") model = AutoModelForCausalLM.from_pretrained("google/gemma-3-270m") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use google/gemma-3-270m with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "google/gemma-3-270m" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "google/gemma-3-270m", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/google/gemma-3-270m
- SGLang
How to use google/gemma-3-270m 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 "google/gemma-3-270m" \ --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": "google/gemma-3-270m", "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 "google/gemma-3-270m" \ --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": "google/gemma-3-270m", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use google/gemma-3-270m with Docker Model Runner:
docker model run hf.co/google/gemma-3-270m
_sliding_window_pattern
Hello. Why the underscore at the beginning of this config setting, making it inconsistent with the previous Gemma 3 model config.json files?
Is it being deprecated and is the layer_types field to be used instead?
Layer_type is the new way to go about it yes!
Hi ,
Thanks for your question! The underscore prefix in config settings often indicates an internal or private attribute intended to avoid conflicts or signal that it’s not part of the public API. This can happen when evolving model configs for better modularity or backward compatibility.
Regarding the google/gemma-3-270m model, it’s possible that this underscore-prefixed setting reflects an internal refactor or interim step as the config format evolves. The layer_types field you mentioned is indeed part of the newer config conventions and might be the preferred way going forward.
For information kindly refer this model’s documentation.
Thank you.