shiyawei comments

Results 4 comments of


                                            shiyawei

get "invalid device context" or "segmentation fault" problems in aarch64 machine

I meet the same problem, can anyone help

分配8G 但是实际使用不会限制在8G内

![gpu_mem_no_limit](https://github.com/AliyunContainerService/gpushare-scheduler-extender/assets/17153725/2fcfae21-46a6-4d0f-8166-a81155feeaf4) 实际使用没有被限制住

errors while in finetune intermlm2-chat-20b with qlora

telechat-12b the same error 09/25/2024 07:48:54 - WARNING - llamafactory.model.model_utils.checkpointing - You are using the old GC format, some features (e.g. BAdam) will be invalid. 09/25/2024 07:48:54 - INFO -...

torch:1.12.0+cu113 驱动:530.41.03，gpu manager调度成功之后，使用cuda报错

针对问题1，也可以尝试降低节点上nv卡的驱动版本及含带的cuda版本，例如 ![image](https://github.com/tkestack/vcuda-controller/assets/17153725/8ba7fb1d-9487-4273-9653-ad190f38ce33) 然后进入业务pod中，执行nvidia-smi命令，查看 ![image](https://github.com/tkestack/vcuda-controller/assets/17153725/b2f710c6-f2d9-4ed2-b1ee-4a527c43bdd4) function not found就会消失，但总显存不是pod分配的显存，还需要解决这个问题