Le Zhang

https://lezhang7.github.io/

I'm a Ph.D student at MILA - Quebec AI Institute, my research focus includes representation learning, vision&language learning

Results 5 comments of


                                            Le Zhang

NCCL broken pipe

same issue here

Low GPU Utilization, help on training parameters. 训练时显卡利用率很低，如何提升训练效率？

这应该不是一个bug, 最后用asycio重写了奖励函数的实现就快了

Would VLLM be used to accelerate model inference?

Does this require implement generate_inner in the clent model by ourselfs?

Question about Cambrian-Alignment dataset

Also there is llava-pretrain and sbu558k both exist in the aligment data, I wonder the difference between them.

Question about Cambrian-Alignment dataset

Thanks, that makes sense then.