alphaedge-ai/Qwen3-1.7B-ast-32768
The alphaedge-ai/Qwen3-1.7B-ast-32768 model is a 1.54 billion parameter variant of the Qwen3-1.7B architecture, specifically optimized for the Asturian language. This model achieves a 24.02% reduction in model size and a 78.43% reduction in vocabulary size compared to its base model by using a trimming method. It maintains a 32,768 token context length and is designed for efficient natural language processing tasks in Asturian, though its performance may be limited for other languages due to vocabulary reduction.
Loading preview...
Overview
alphaedge-ai/Qwen3-1.7B-ast-32768 is a specialized language model derived from the Qwen3-1.7B architecture, featuring 1.54 billion parameters and a 32,768 token context length. Its primary distinction lies in its optimization for the Asturian language through a vocabulary trimming process. This method significantly reduces the model's memory footprint and vocabulary size while aiming to maintain similar performance to the original Qwen3-1.7B for its target language.
Key Characteristics
- Language Specialization: Optimized specifically for the Asturian language.
- Efficiency: Achieves a 24.02% reduction in model size (from 2.03B to 1.54B parameters) and a 78.43% reduction in vocabulary size (from 151,936 to 32,768 tokens) compared to the base Qwen3-1.7B model.
- Trimming Method: Utilizes a vocabulary trimming technique to remove tokens not commonly used in Asturian, enhancing efficiency for the target language.
- Context Length: Retains a substantial context window of 32,768 tokens.
Intended Use Cases
- Asturian NLP Applications: Ideal for tasks requiring natural language understanding and generation in Asturian.
- Resource-Constrained Environments: Suitable for deployments where a smaller model size and reduced memory usage are critical, provided the focus is on Asturian.
Limitations
- Language Specificity: Due to the vocabulary trimming, this model may exhibit reduced performance for languages other than Asturian.