Parum-Lucis/gemma3-270m-bpe-4K-5y-merged

TEXT GENERATIONConcurrent Unit Cost:1Model Size:0.3BQuant:BF16Context Size:32kPublished:Jul 8, 2026Architecture:Transformer Featherless Exclusive Cold

Parum-Lucis/gemma3-270m-bpe-4K-5y-merged is a 0.3 billion parameter language model based on the Google Gemma-3 architecture. This model was created using the Task Arithmetic merge method, combining the base google/gemma-3-270m with rafurafu/gemma-3-270m-cpt-bpe-4K-5y. It is designed for efficient deployment in resource-constrained environments, leveraging its compact size and merged architecture for specific applications.

Loading preview...

Model Overview

This model, gemma3-270m-bpe-4K-5y-merged, is a compact language model with approximately 0.3 billion parameters, built upon the Google Gemma-3 architecture. It was developed by Parum-Lucis through a merging process using mergekit.

Merge Details

The model was created using the Task Arithmetic merge method, which combines the strengths of different pre-trained models. The base model for this merge was google/gemma-3-270m. It was merged with rafurafu/gemma-3-270m-cpt-bpe-4K-5y to potentially enhance its capabilities for specific tasks or data distributions.

Key Characteristics

  • Architecture: Based on the Gemma-3 family from Google.
  • Parameter Count: 0.3 billion parameters, making it suitable for efficient inference.
  • Merge Method: Utilizes Task Arithmetic for combining model weights.
  • Context Length: The underlying Gemma-3 architecture typically supports a context length of 32768 tokens.

Use Cases

This model is particularly well-suited for applications requiring a small, efficient language model. Its merged nature suggests potential optimizations for:

  • Edge device deployment: Due to its compact size.
  • Specific domain tasks: Where the merged components might offer specialized knowledge.
  • Rapid prototyping and experimentation: For quick iteration cycles.