gouki510/gemma2-2b-base-filtered-geometry-keep-0.3

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:2.6BQuant:BF16Context Size:8kPublished:Nov 18, 2025License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The gouki510/gemma2-2b-base-filtered-geometry-keep-0.3 model is a 2.6 billion parameter Gemma2-based language model developed by gouki510, fine-tuned using Unsloth and Huggingface's TRL library. This model was specifically trained to be twice as fast as its base counterpart. With an 8192-token context length, it offers efficient processing for various language tasks.

Loading preview...

Overview

The gouki510/gemma2-2b-base-filtered-geometry-keep-0.3 is a 2.6 billion parameter language model, developed by gouki510. It is fine-tuned from the unsloth/gemma-2-2b base model, leveraging the Unsloth library and Huggingface's TRL library for its training process. A key characteristic of this model is its optimized training, which resulted in a 2x speed improvement compared to the original Gemma2-2B model.

Key Capabilities

  • Efficient Performance: Achieves significantly faster training times due to the use of Unsloth.
  • Gemma2 Architecture: Benefits from the underlying capabilities of the Gemma2 model family.
  • Context Length: Supports an 8192-token context window, suitable for processing moderately long inputs.

Good For

  • Developers seeking a Gemma2-based model with enhanced training efficiency.
  • Applications where faster fine-tuning and deployment are critical.
  • General language understanding and generation tasks that can benefit from a 2.6 billion parameter model with an 8K context.