Thank you for you amazing work of Ctrl-Adapter! I was trying to run the code on a single NVIDIA 3090 GPU, but I came into the OOM error. Could you please enlighten me what GPU resources do you use to inference the overall pipeline? Is there any efficient solutions to inference Ctrl-Adapter on a single NVIDIA 3090 GPU?
Besides, I have tried changing .to(device) into .enable_model_cpu_offload() (a function in the accelerate library). But the code came into error somewhere in the middle of pipline.
Thank you in advance for your precious time and attention from your busy schedule.
Hi @HL-hanlin ,
Thank you for you amazing work of Ctrl-Adapter! I was trying to run the code on a single NVIDIA 3090 GPU, but I came into the OOM error. Could you please enlighten me what GPU resources do you use to inference the overall pipeline? Is there any efficient solutions to inference Ctrl-Adapter on a single NVIDIA 3090 GPU?
Besides, I have tried changing
.to(device)
into.enable_model_cpu_offload()
(a function in theaccelerate
library). But the code came into error somewhere in the middle of pipline.Thank you in advance for your precious time and attention from your busy schedule.
Best regards