Yeah, the dream is to have a "reasoning kernel" that you can plug specific knowledge and skills into ("I know kung fu"). I'm not aware of any progress on the dream though. Plenty of people are making LoRAs for general-purpose open models and there are companies like Thinking Machines that will help you. I'm not sure how you'd choose which base model to start with. I'm also not sure how to draw the boundary between LoRA and RAG.
Comments
Yeah, the dream is to have a "reasoning kernel" that you can plug specific knowledge and skills into ("I know kung fu"). I'm not aware of any progress on the dream though. Plenty of people are making LoRAs for general-purpose open models and there are companies like Thinking Machines that will help you. I'm not sure how you'd choose which base model to start with. I'm also not sure how to draw the boundary between LoRA and RAG.
Some links I found but I can't vouch for them:
https://github.com/oobabooga/textgen/wiki/05-%E2%80%90-Train...
https://thinkingmachines.ai/tinker/