amphora/Qwen3-4B-32K-PLZPLZ
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:May 28, 2026License:apache-2.0Architecture:Transformer Open Weights Warm
amphora/Qwen3-4B-32K-PLZPLZ is a 4 billion parameter Qwen3 model developed by amphora, featuring a 32K context length. This model was fine-tuned from amphora/Qwen3-4B-DASD-32K and optimized for training speed using Unsloth. Its primary differentiator is its accelerated training process, making it suitable for applications requiring efficient model development.
Loading preview...
Model Overview
amphora/Qwen3-4B-32K-PLZPLZ is a 4 billion parameter Qwen3 model, developed by amphora, that supports a substantial 32K token context length. This model is a fine-tuned variant of amphora/Qwen3-4B-DASD-32K.
Key Characteristics
- Architecture: Based on the Qwen3 model family.
- Parameter Count: 4 billion parameters, offering a balance between performance and computational efficiency.
- Context Length: Features a 32,768 token context window, enabling processing of extensive inputs and generating longer, more coherent outputs.
- Training Optimization: A significant differentiator is its training process, which was accelerated by 2x using the Unsloth library. This optimization focuses on enhancing training efficiency.
Potential Use Cases
- Efficient Development: Ideal for developers and researchers looking for a Qwen3-based model that can be fine-tuned or adapted more rapidly due to its optimized training heritage.
- Long Context Applications: Suitable for tasks requiring understanding and generation over long documents, conversations, or codebases, thanks to its 32K context window.
- Resource-Conscious Deployment: As a 4B parameter model, it offers a more accessible option for deployment compared to larger models, while still providing robust language capabilities.