SEED
SEED copied to clipboard
Request for Instruction Tunning
I've noticed you mentioned interleaved text-image generation only able on instruction-tuned model, would you release in the future? or would you please release some methods of instruction-tuning?