Model Overview
This model, PetarKal/Qwen3-4B-Base-ascii-art-v5-e3-lr5e-6-ga16-ctx4096, is a specialized 4 billion parameter language model. It is a fine-tuned variant of the Qwen/Qwen3-4B-Base architecture, developed by PetarKal. The fine-tuning process utilized the TRL (Transformers Reinforcement Learning) framework, specifically employing Supervised Fine-Tuning (SFT).
Key Capabilities
- Specialized ASCII Art Generation: The model's primary and unique capability is generating ASCII art, distinguishing it from general-purpose language models.
- Qwen3-4B-Base Foundation: Built upon the robust Qwen3-4B-Base model, providing a strong linguistic understanding base.
- Extended Context Length: Supports a context length of 32768 tokens, allowing for more intricate and detailed ASCII art creations or longer input prompts.
Training Details
The model was trained using SFT (Supervised Fine-Tuning) with TRL version 0.29.1, Transformers 5.3.0, Pytorch 2.5.1, Datasets 4.8.4, and Tokenizers 0.22.2. Further details on the training run can be visualized via Weights & Biases.
When to Use This Model
This model is ideal for applications requiring creative text-to-ASCII art conversion. If your use case involves generating stylized text, visual representations using characters, or integrating ASCII art into creative projects, this model offers a specialized solution. It is not intended for general conversational AI, code generation, or factual question answering, but rather excels in its niche domain of ASCII art production.