kevinkawchak/nvidia-Llama3-ChatQA-1.5-8B-Molecule16
kevinkawchak/nvidia-Llama3-ChatQA-1.5-8B-Molecule16 is an 8 billion parameter language model, fine-tuned by kevinkawchak from nvidia/Llama3-ChatQA-1.5-8B. It specializes in molecule-oriented instruction following, particularly for description-guided molecule design, using the zjunlp/Mol-Instructions dataset. This model is designed to answer general questions and biochemistry-related questions, providing SELFIES structures for molecular compounds.
Loading preview...
Overview
kevinkawchak/nvidia-Llama3-ChatQA-1.5-8B-Molecule16 is an 8 billion parameter language model, fine-tuned from NVIDIA's Llama3-ChatQA-1.5-8B. This model specializes in processing and generating information related to molecules, particularly for description-guided molecule design. It leverages the zjunlp/Mol-Instructions dataset, which focuses on molecule-oriented instructions.
Key Capabilities
- Molecule-Oriented Instruction Following: Excels at understanding and responding to prompts related to molecular structures and design.
- Biochemistry Question Answering: Capable of answering general and biochemistry-specific questions, often returning molecular structures in SELFIES format.
- Fine-tuned for Specificity: Built upon a base model known for enhancing tabular and arithmetic calculation, this version is further specialized for chemical information.
Good for
- Drug Discovery Research: Assisting in early-stage drug discovery by generating or interpreting molecular descriptions.
- Chemical Informatics: Applications requiring the conversion of descriptive text into structured chemical representations (SELFIES).
- Educational Tools: Providing structured answers to biochemistry and chemistry questions, particularly those involving molecular structures.