Casual132/gemma-3-1b-finetuned-lora-loss3.9

TEXT GENERATIONConcurrency Cost:1Model Size:1BQuant:BF16Ctx Length:32kPublished:Apr 26, 2026License:gemmaArchitecture:Transformer Cold

Casual132/gemma-3-1b-finetuned-lora-loss3.9 is a 1 billion parameter multimodal model from the Gemma 3 family by Google DeepMind, built from the same technology as Gemini models. This variant is fine-tuned with LoRA and features a 32K token context window, handling both text and image inputs to generate text outputs. It excels at diverse text generation and image understanding tasks, including question answering, summarization, and reasoning, while being optimized for deployment in resource-limited environments.

Loading preview...

Model Overview

Casual132/gemma-3-1b-finetuned-lora-loss3.9 is a 1 billion parameter model from the Gemma 3 family, developed by Google DeepMind. These models are lightweight, multimodal, and share technology with the Gemini models. This specific variant is fine-tuned using LoRA and is designed for efficient deployment.

Key Capabilities

  • Multimodal Input: Processes both text and image inputs. Images are normalized to 896x896 resolution and encoded to 256 tokens.
  • Text Generation: Generates text outputs for tasks like question answering, summarization, and creative content creation.
  • Image Understanding: Capable of analyzing image content and extracting visual data.
  • Context Window: Features a 32K token input context window, suitable for various tasks.
  • Multilingual Support: Trained on data including over 140 languages.

Training and Performance

The 1B model was trained on 2 trillion tokens, encompassing web documents, code, mathematics, and images. It demonstrates strong performance across reasoning, STEM, code, and multilingual benchmarks for its size. For instance, it achieves 62.3 on HellaSwag (10-shot) and 73.0 on ARC-e (0-shot).

Intended Usage

This model is well-suited for applications requiring text generation, chatbots, text summarization, and image data extraction. Its relatively small size makes it ideal for deployment in environments with limited resources, such as laptops or edge devices, democratizing access to advanced AI capabilities.