danish-foundation-models/gemma-3-1b-scratch-dynaword-full-v1
The danish-foundation-models/gemma-3-1b-scratch-dynaword-full-v1 is a 1 billion parameter language model, part of the Gemma family, developed by danish-foundation-models. It was specifically trained as part of the Dynaword paper to demonstrate improvements from training on the Danish Dynaword dataset. This model is optimized for tasks requiring understanding and generation of Danish language, leveraging its 32768 token context length.
Loading preview...
Model Overview
The danish-foundation-models/gemma-3-1b-scratch-dynaword-full-v1 is a 1 billion parameter model from the Gemma architecture, developed by danish-foundation-models. It was created as part of the research for the Dynaword paper.
Key Characteristics
- Architecture: Based on the Gemma model family.
- Parameter Count: Features 1 billion parameters, offering a balance between performance and computational efficiency.
- Context Length: Supports a substantial context window of 32768 tokens.
- Training Focus: Specifically trained to demonstrate the relative improvement achieved by utilizing the Danish Dynaword dataset.
Intended Use Cases
This model is particularly well-suited for research and applications focused on:
- Danish Language Processing: Ideal for tasks requiring robust understanding and generation in Danish, given its specialized training on the Danish Dynaword dataset.
- Experimental Research: Useful for researchers exploring the impact of dynamic dataset development on model performance, as outlined in the Dynaword paper.