juvereturn/Qwen3-1.7B-CS592-Final
The juvereturn/Qwen3-1.7B-CS592-Final is a 2 billion parameter language model, fine-tuned from the Qwen3-1.7B architecture using QLoRA. Developed by juvereturn, this model is specifically designed as a Bible assistant, capable of answering questions about the Bible. It is optimized for single-turn interactions, serving as a specialized tool for biblical queries.
Loading preview...
Model Overview
The juvereturn/Qwen3-1.7B-CS592-Final is a 2 billion parameter language model, fine-tuned from the base Qwen/Qwen3-1.7B architecture. This model was developed by juvereturn as a test model for the CS-394/594 class at DigiPen, utilizing QLoRA (4-bit) with supervised fine-tuning.
Key Capabilities
- Bible Assistant: The primary function of this model is to act as a Bible assistant, providing answers to questions related to the Bible.
- Emotional Support: It is intended to help users feel better through its biblical responses.
- Single-Turn Interactions: The model is designed and optimized for single-turn conversations, meaning it processes one query at a time without maintaining conversational context over multiple turns.
Training Details
The model was fine-tuned using the juvereturn/CS592-Final dataset. The training involved 3 epochs with a learning rate of 0.0002, utilizing LoRA with a rank of 16 and an alpha value of 32.
Limitations
It is important to note that this model is a single-turn model and has not been trained to support long, multi-turn conversations. Its performance is best suited for direct, one-off questions.