HashNuke/google-gemma-3-270m-it

TEXT GENERATIONConcurrent Unit Cost:1Model Size:0.3BQuant:BF16Context Size:32kPublished:Nov 24, 2025License:gemmaArchitecture:Transformer Featherless Exclusive Cold

HashNuke/google-gemma-3-270m-it is a 0.3 billion parameter instruction-tuned multimodal model from Google DeepMind, part of the Gemma 3 family. Built from the same research as Gemini models, it handles text and image input to generate text output, supporting over 140 languages. This lightweight model is optimized for a variety of text generation and image understanding tasks, including question answering and summarization, and is suitable for resource-constrained environments.

Loading preview...

Model Overview

HashNuke/google-gemma-3-270m-it is a 0.3 billion parameter instruction-tuned model from Google DeepMind, belonging to the Gemma 3 family. These models are lightweight, multimodal, and built using the same research and technology as the Gemini models. They are capable of processing both text and image inputs to generate text outputs, with open weights available for both pre-trained and instruction-tuned variants. The 270M model has a 32K token context window and supports over 140 languages.

Key Capabilities

  • Multimodal Input: Accepts text strings and images (normalized to 896x896 resolution, encoded to 256 tokens each for larger Gemma 3 models, though the 270M is primarily text-focused with image capability in the family).
  • Text Generation: Excels at generating creative text formats, chatbot responses, and text summarization.
  • Image Understanding: Can extract, interpret, and summarize visual data for text communications.
  • Multilingual Support: Trained on data including content in over 140 languages.
  • Resource-Efficient: Its relatively small size makes it suitable for deployment on devices with limited resources like laptops or desktops.

Use Cases

  • Content Creation: Generating various text formats, marketing copy, or email drafts.
  • Conversational AI: Powering chatbots and virtual assistants.
  • Research & Education: Serving as a foundation for VLM/NLP research, language learning tools, and knowledge exploration.
  • Image Data Extraction: Analyzing image content and summarizing visual information.

Finetunes
1 models