PetarKal/Qwen3-4B-Base-ascii-art-v6-phase2b-generation-lr1e5
PetarKal/Qwen3-4B-Base-ascii-art-v6-phase2b-generation-lr1e5 is a 4 billion parameter Qwen3-based language model fine-tuned by PetarKal. This model is a fine-tuned version of a previous phase focused on understanding ASCII art, now specialized for text generation. It was trained using the TRL framework and is designed for generating responses based on given prompts, building upon its ASCII art understanding capabilities.
Loading preview...
Model Overview
This model, PetarKal/Qwen3-4B-Base-ascii-art-v6-phase2b-generation-lr1e5, is a 4 billion parameter language model built on the Qwen3 architecture. It represents the second phase of fine-tuning, specifically focusing on text generation. This model builds upon its predecessor, PetarKal/Qwen3-4B-Base-ascii-art-v6-phase1-understanding, which was designed for understanding ASCII art.
Key Capabilities
- Text Generation: Optimized for generating coherent and relevant text based on user prompts.
- Fine-tuned Qwen3 Base: Leverages the foundational capabilities of the Qwen3-4B-Base model.
- TRL Framework: Trained using the Transformers Reinforcement Learning (TRL) library, indicating a focus on improving generation quality through advanced fine-tuning techniques.
Training Details
The model underwent Supervised Fine-Tuning (SFT) as part of its training procedure. It was developed using specific versions of key frameworks:
- TRL: 0.29.1
- Transformers: 5.5.0
- Pytorch: 2.10.0
- Datasets: 4.8.4
- Tokenizers: 0.22.2
Intended Use
This model is suitable for applications requiring text generation, particularly in contexts where a foundation in understanding ASCII art (from its phase 1 predecessor) might be beneficial for specific generation tasks. Developers can integrate it using the Hugging Face pipeline for text generation tasks.