gcxamy

Results 1 comments of gcxamy

> I train the LoRA model on 4 GPUs, fp16 precision. Other options are left as default. I haven't tried libraries like deepspeed, feel free to share your experience if...