thuml / TimesNet

About Code release for "TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis" (ICLR 2023), https://openreview.net/pdf?id=ju_Uqw384Oq

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

GPU显存不足

lishijie15 opened this issue · comments

同学你好,在运行时碰到了以下的问题
RuntimeError: CUDA out of memory. Tried to allocate 288.00 MiB (GPU 0; 23.99 GiB total capacity; 22.60 GiB already allocated; 0 bytes free; 22.75 GiB reserved in total by PyTorch)
在运行timesnet前我确定显存已经清理干净了,也尝试调小batch_size,但是仍然超显存,你论文中的显卡也是24g的,我想问问你是否有遇到这样的问题,你是怎样处理的呢?

您好,我也是24GB显存的机器,而且timesnet所有的实验都不会超过20GB,具体是跑哪个script出问题的呢?

您好,运行参数为--task_name long_term_forecast --is_training 1 --model_id train --model TimesNet --data ETTh2

image

commented

似乎您没有设置d_model和d_ff,run.py里面的默认值比较大。建议使用我们提供的script跑,具体为:
sh ./scripts/long_term_forecast/ETT_script/TimesNet_ETTh2.sh

好的,已经解决了,谢谢