SeongryongJung/Qwen3-4B-Chemical-RLSD-TR
SeongryongJung/Qwen3-4B-Chemical-RLSD-TR is a 4 billion parameter Qwen3-based language model fine-tuned using the RLSD_TR method specifically for chemical domain tasks. This model demonstrates specialized performance on the Chemical / SciKnowEval chemistry dataset, achieving a validation mean@16 score of 68.96%. It is optimized for applications requiring chemical knowledge and reasoning, leveraging a 32768 token context length.
Loading preview...
Model Overview
SeongryongJung/Qwen3-4B-Chemical-RLSD-TR is a specialized 4 billion parameter language model built upon the Qwen3-4B architecture. It has been fine-tuned using the RLSD_TR (Reinforcement Learning from Simulated Data with Trust-Region regularization) method, specifically targeting the chemical domain.
Key Capabilities & Performance
- Chemical Domain Expertise: This model is explicitly trained on the
Chemical / SciKnowEval chemistrydataset, indicating a strong focus on chemical knowledge and problem-solving. - Performance: Achieved a validation mean@16 score of 68.96% on the chemistry dataset after 100 training steps, demonstrating its proficiency in the targeted domain.
- Training Method: Utilizes the
RLSD_TRmethod with a batch size of 32, a learning rate of1e-6, and a trust-region mix/teacher update rate of0.1. - Context Length: Supports a maximum prompt length of 2048 tokens and a maximum response length of 8192 tokens, with a total model length capacity of 10240 tokens.
Use Cases
This model is particularly well-suited for applications requiring a deep understanding of chemistry, such as:
- Answering chemistry-related questions.
- Assisting with chemical research and data analysis.
- Generating or interpreting chemical information.
Developers can integrate this model using the Hugging Face transformers library for tasks within the chemical domain.