Tom Young

icon indicating a website link https://tomyoung903.github.io/

icon company NTU icon location 50 Nanyang Ave, 639798 icon location Tom Young is a research fellow at NUS working on language models.

Results 7 comments of Tom Young

How to use the code for multinode inference?

Hi thanks for the response. But I guess I didn't make myself clear. I have enough gpus so no quantization is needed for me. However, those gpus exist among multiple...

How to use the code for multinode inference?

much thanks. but I did it with just torchrun with some rdzv arguments

ZeroDivisionError: division by zero

GPT2 is no longer the state of the art backbone. I would suggest instruction tuning + large language models

preprocessing details?

thank you Cheers, Tom Young PhD Student School of Computer Science and Engineering (SCSE) Nanyang Technological University (NTU) 50 Nanyang Ave, Singapore 639798 tomyoung903.github.io On Mon, Apr 19, 2021 at...

[BUG]: ModuleNotFoundError: No module named 'colossalai._C.cpu_adam'

Hi have you found a solution to this problem? I am encountering the same problem with colossalai-0.3.2, torch.2.2.0.dev+cu121 ,cuda12.2

[BUG]: ModuleNotFoundError: No module named 'colossalai._C.cpu_adam'

Thanks! I heard colossal-ai was tested on h800. What env (cuda, torch) was used?

Need to reorder the flow tensor when transpose_image is true?

No it's not about transposing the flow tensor. It's about reordering the dimension that has size 2 flows[:, [0,1]] = flows[:, [1,0]] such that the first slice always corresponds to...