kristine-li

problem of training

1

when I follow the step of training, the results show that:  when epoch = 240 ,mAP is also very low, about 0.02 (batch-saize, lr are defalut). can you give...

QLoRA微调后模型推理过程中模型加载失败的情况

4

 CUDA_VISIBLE_DEVICES=1 python cli_demo.py --from_pretrained ./checkpoints/finetune-visualglm-6b-06-01-06-47/