Srishtik/gemma-3-270m-linear-merged-trained-on-responses

TEXT GENERATIONConcurrent Unit Cost:1Model Size:0.3BQuant:BF16Context Size:32kPublished:Jun 12, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

Srishtik/gemma-3-270m-linear-merged-trained-on-responses is a 0.3 billion parameter Gemma-3 based language model developed by Srishtik. Fine-tuned from unsloth/gemma-3-270m-it, this model was trained significantly faster using Unsloth. It is designed for tasks involving text generation and understanding, leveraging its efficient training for responsive applications.

Loading preview...

Model Overview

Srishtik/gemma-3-270m-linear-merged-trained-on-responses is a compact 0.3 billion parameter language model, developed by Srishtik. It is fine-tuned from the unsloth/gemma-3-270m-it base model, indicating its foundation in the Gemma-3 architecture.

Key Characteristics

  • Efficient Training: This model was trained approximately two times faster by leveraging the Unsloth library. This suggests an optimization for training speed and resource efficiency.
  • Base Model: It builds upon the gemma-3-270m-it model, implying a focus on instruction-following capabilities inherent to its base.
  • License: The model is released under the Apache-2.0 license, providing permissive use for developers.

Potential Use Cases

Given its efficient training and Gemma-3 base, this model is suitable for:

  • Applications requiring a lightweight and fast-to-deploy language model.
  • Tasks where rapid iteration and fine-tuning are beneficial.
  • Instruction-following tasks in resource-constrained environments.