iamzycon

I'm considering setting up a fine-tuning and inference platform for Llama that would allow customers to host their fine-tuned models. Would it be necessary to allocate a dedicated infrastructure for each fine-tuned model, or could a shared infrastructure work? Are there any existing solutions for this?