Skip to main content

Gemma3 270M IT

A compact 270M parameter multimodal language model from Google optimized for text classification and data extraction tasks with rapid fine-tuning capability.

Supported processors

  • SL2619

Performance

Torq Compiler/Runtime v2.3.0

1668ms
Time to First Token
8.3tok/s
Tokens per Second

Gemma3 270M

Model Overview​

The Gemma 3 270M is a highly compact, 270-million parameter multimodal model from Google designed for extreme efficiency. It is specifically engineered for well-defined tasks like text classification and data extraction where it offers a cost-effective alternative to larger models. Its lightweight architecture is particularly optimized for rapid fine-tuning.

Gemma3 270M been optimized for the Synaptics Astra® SL2610-Series processors with Torq NPU.

Model Features​

  • Model Type: Large Language Model
  • Input: Text
  • Output: Text

Deployment​

The compiled model files are available for download on Huggingface at Synaptics/Gemma-3-270m-it.

Usage Tutorial​

A usage example is provided at Torq Examples / gemma3. Check out the README for instructions.

License​

Both the source model and the compiled model for on-device deployment are licensed under Gemma Terms of Use.

Learn More​

Related models