google

/

gemma-3-270m

like 1.02k

Follow

Google 55.4k

Hai

Hi @Najin06 ,

That's great! The Gemma 3 270M model is an excellent choice for local experimentation because of its small size.

please follow the below step by step instructions :

1. Set up environment :
   First, make sure you have Python (>=3.8) installed, along with pip. Then, set up a virtual environment.

2. Install Required Libraries :
   Install Hugging Face Transformers and PyTorch (or TensorFlow)

3. Download the Model :
   Use the Hugging Face transformers library to load the model:

from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "google/gemma-3-270m"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

This will automatically download the model weights and tokenizer to your local machine.

4. Run a Simple Inference :
   input_text = "What is the capital of France?"
inputs = tokenizer(input_text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=50)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

For experimentation, consider enabling torch.no_grad() during inference to save memory.

Kindly follow this steps and let us know if you have any concerns will assist you on this.

Thank you.