Currently, the repository supports a limited number of base models for conversion from Decoder-only LLM to Embedding models. To make this tool more widely applicable and valuable to the community, we should expand support to include more popular and diverse base models.