Comment on QLoRA: Efficient Finetuning of Quantized LLMsparentComments−zirgs3yWe could try something like Civitai is doing already, but automated.Each node could train the model on a separate concept and then combine the results.
Comments
We could try something like Civitai is doing already, but automated.
Each node could train the model on a separate concept and then combine the results.