tncyypttvk/Prom-1.5B
Prom-1.5B is a 1.5 billion parameter instruction-tuned causal language model developed by tncyypttvk, based on the Qwen2.5-1.5B-Instruct architecture, with a context length of 32768 tokens. It is specifically fine-tuned as a bilingual (Turkish and English) student affairs assistant for a fictional university, excelling in tasks like identity recognition, RAG-based information retrieval, and tool-use for GPA calculation. Its primary use case is providing automated support for university-related queries in both languages.
Loading preview...
Prom-1.5B: Bilingual Student Affairs Assistant
Prom-1.5B is a 1.5 billion parameter language model developed by tncyypttvk, fine-tuned from the Qwen/Qwen2.5-1.5B-Instruct base model. Designed as a bilingual (Turkish and English) student affairs assistant for a fictional university, it leverages a small model size for efficient deployment on T4 GPUs and CPU inference. The model was trained using QLoRA on 648 synthetically generated chat conversations covering identity, RAG, and agent-based tasks, with a focus on university-specific information.
Key Capabilities
- Bilingual Support: Handles queries in both Turkish and English, acting as a student affairs assistant.
- Information Retrieval (RAG): Capable of retrieving information from provided guides and course data.
- Tool Use: Integrates with a custom tool for GPA calculation based on entered course grades.
- Identity Recognition: Demonstrates improved performance in recognizing specific identities in both languages.
- Efficient Deployment: Its 1.5B parameter size makes it suitable for resource-constrained environments.
Good for
- Developing small-scale, specialized chatbots for educational institutions or similar domains.
- Bilingual conversational AI applications requiring support in Turkish and English.
- Experimenting with tool-augmented LLMs for structured data interaction (e.g., GPA calculation).
- Use cases where efficient inference on CPU or smaller GPUs is a priority.