User:honeyzfxy213906

You might have to make use of the gpu_memory_limit and/or lora_on_cpu config options to prevent working outside of memory. If you still run from CUDA memory, you may make an effort to merge in method

https://imtoken-wt.com/