METR-LA

Dataset Information
Modalities
Time series
License
Unknown
Contents

Overview

METR-LA is a dataset for traffic prediction.

Associated Benchmarks

This dataset is used in 1 benchmark:

  • Traffic Prediction - Metrics: MAE @ 12 step, 12 steps MAE, 12 steps MAPE, 12 steps RMSE, MAE @ 3 step

Recent Benchmark Submissions

Task Model Paper Date
Traffic Prediction T-Graphormer T-Graphormer: Using Transformers for Spatiotemporal … 2025-01-22
Traffic Prediction TITAN A Time Series is Worth … 2024-09-26
Traffic Prediction STD-MAE Spatial-Temporal-Decoupled Masked Pre-training for Spatiotemporal … 2023-12-01
Traffic Prediction STAEformer STAEformer: Spatio-Temporal Adaptive Embedding Makes … 2023-08-21
Traffic Prediction DCGCN Dynamic Causal Graph Convolutional Network … 2023-06-12
Traffic Prediction MegaCRN Spatio-Temporal Meta-Graph Learning for Traffic … 2022-11-27
Traffic Prediction STEP Pre-training Enhanced Spatial-temporal Graph Neural … 2022-06-18
Traffic Prediction D2STGNN Decoupled Dynamic Spatial-Temporal Graph Neural … 2022-06-18
Traffic Prediction ADN Structured Time Series Prediction without … 2022-02-07
Traffic Prediction Finetune from t1-6 checkpoint Incrementally Improving Graph WaveNet Performance … 2019-12-11
Traffic Prediction Graph WaveNet Graph WaveNet for Deep Spatial-Temporal … 2019-05-31
Traffic Prediction ST-UNet ST-UNet: A Spatio-Temporal U-Network for … 2019-03-13
Traffic Prediction STGCN Spatio-Temporal Graph Convolutional Networks: A … 2017-09-14
Traffic Prediction DCRNN Diffusion Convolutional Recurrent Neural Network: … 2017-07-06

Research Papers

Recent papers with results on this dataset:

External Links: