Local LLM

[justantopair-ai]

Between gemma-9b and llama-8b, which one should be better for JAV translation? do we have any future plan for bigger model for 12/16 GB VRAM, or a way to import custom LLM?
For specific categories like

• drama
• decent amount of dialogue
• training
• blackmail
• hypnosis
• whisper suggestion (kinda similar to ASMR)

and would balanced + aggressive better in this use case or stick with balanced + balanced?