tongnie / nexusqn

Repository for ''Contextualizing MLP-Mixers Spatiotemporally for Urban Data Forecast at Scale''

Repository for ''Contextualizing MLP-Mixers Spatiotemporally for Urban Data Forecast at Scale''

Code structure:

• Model:
  • nexusqn_model.py
  • mlp.py
  • embedding.py
• Config:
  • traffic:
    • default
    • nexusqn_pems08.yaml

main file: run_traffic_benchmark.py

The appendix can be found at urban_forecast_appendix.pdf

More information can be found in our preprint.

Acknowledgement

Our code is built upon the TorchSpatiotemporal repository (https://github.com/TorchSpatiotemporal/tsl).

nexusqn's Issues

How to run run_traffic_benchmark.py?

Hello, torch-spatiotemporal What version are you using? Can you provide the "requirement" file?
I can't execute the run_traffic_benchmark.py script. The mistake is module 'tsl.nn.models' has no attribute 'NexuSQNModel', where should I put the nexusqn_model file.
Hope to get your answer and guidance, thank you!