Thinklab-SJTU/Crossformer
Official implementation of our ICLR 2023 paper "Crossformer: Transformer Utilizing Cross-Dimension Dependency for Multivariate Time Series Forecasting"
PythonApache-2.0
Issues
- 2
关于超参数调节
#29 opened by Certainguys - 1
Changing routers
#26 opened by abase123 - 1
- 0
关于只想预测个别向量
#28 opened by sdGundamf - 2
Tools used to create model illustrations
#27 opened by FishCatCake - 4
Info
#20 opened by georgesmiaka - 1
- 1
- 1
- 1
关于MTGNN
#21 opened by teethdocker - 1
关于预测数据不准确的问题
#15 opened by lsl0000 - 1
多变量预测单变量具体怎么修改
#9 opened by smallkaka - 6
- 3
attn.py中的mix?
#12 opened by Blues-He - 0
请问为什么在解码器阶段要将不同尺度的预测结果相加呢
#19 opened by VuGum - 4
过拟合严重
#18 opened by lilpeng-xxxl - 2
不同channel使用同一个linear层做projection的效果
#17 opened by ligrire - 0
inverse设置成true后loss值特别大
#16 opened by ICBC0921 - 3
train loss 比 valid和test loss大是什么原因?
#14 opened by joezou - 2
- 2
关于SegMerging的请教
#11 opened by joeySJTU - 1
关于跨维度阶段的注意力
#8 opened by zixuanxuanxuanxuan - 1
inverse_transform
#3 opened by houzimm - 2
Reproduction issue with weather dataset
#5 opened by lss-1138 - 2
关于weather数据集的问题
#4 opened by vaeaea - 2