MetaTTE: a Meta-Learning Based Travel Time Estimation Model for Multi-city Scenarios

morningstarwang

Last update: Dec 14, 2022

Related tags

Deep Learning MetaTTE

Overview

MetaTTE: a Meta-Learning Based Travel Time Estimation Model for Multi-city Scenarios

This is the official TensorFlow implementation of MetaTTE in the manuscript.

Core Requirements

tensorflow~=2.3.0
numpy~=1.18.4
spektral~=0.6.1
pandas~=1.0.3
tqdm~=4.46.0
opencv-python~=4.3.0.36
matplotlib~=3.2.1
Pillow~=7.1.2
scipy~=1.4.1

All Dependencies can be installed using the following command:

pip install -r requirements.txt

Data Preparation

We here provide the datasets we adopted in this paper with Google Drive. After downloading the zip file, please extract all the files in data directory to the data folder in this project.

Download Link: Download

Configuration

We here list a sample of our config file, and leave the comments for explanation. \ (Please DO NOT include the comments in config files)

[General]
mode = train
# Specify the absoulute path of training, validation and testing files
train_files = ./data/chengdu/train.npy,./data/porto/train.npy
val_files = ./data/chengdu/val.npy,./data/porto/val.npy
test_files = ./data/chengdu/test.npy,./data/porto/test.npy
# Specify the batch size
batch_size = 32
# Specify the number for GPU
gpu = 7
# Specify the unique label for each experiment
prefix = tte_exp_64_gru

[Model]
# Specify the inner learning rate
learning_rate = 1e-2
# Specify the inner reduce rate of learning rate
lr_reduce = 0.5
# Specify the maximum iteration
epoch = 500000
# Specify the k shot
inner_k = 10
# Specify the outer step size
outer_step_size = 0.1
# Specify the model according to the class name
model = MSMTTEGRUAttModel
# Specify the dataset according to the class name
dataset = MyDifferDatasetWithEmbedding
# Specify the dataloader according to the class name
dataloader = MyDataLoaderWithEmbedding


# mean, standard deviation for latitudes, longitudes and travel time (Chengdu is before the comma while Porto is after the comma)
[Statistics]
lat_means = 30.651168872309235,41.16060653954797
lng_means = 104.06000501543934,-8.61946359614912
lat_stds = 0.039222931811691585,0.02315827641949562
lng_stds = 0.045337940910596744,0.029208656457667292
labels_means = 1088.0075248390972,691.2889878452086
labels_stds = 1315.707363003298,347.4765869900725

Model Training

Here are commands for training the model on both Chengdu and Porto tasks.

python main.py --config=./experiments/finetuning/64/gru.conf

Eval baseline methods

Here are commands for testing the model on both Chengdu and Porto tasks.

python main.py --config=./experiments/finetuning/64/gru.conf

Citation

We currently do not provide citations.

CRISCE: Automatically Generating Critical Driving Scenarios From Car Accident Sketches

CRISCE: Automatically Generating Critical Driving Scenarios From Car Accident Sketches This document describes how to install and use CRISCE (CRItical

Chair of Software Engineering II, Uni Passau

2 Feb 9, 2022

Time-series-deep-learning - Developing Deep learning LSTM, BiLSTM models, and NeuralProphet for multi-step time-series forecasting of stock price.

Stock Price Prediction Using Deep Learning Univariate Time Series Predicting stock price using historical data of a company using Neural networks for

7 Nov 27, 2022

This repo is developed for Strong Baseline For Vehicle Re-Identification in Track 2 Ai-City-2021 Challenges

A STRONG BASELINE FOR VEHICLE RE-IDENTIFICATION This paper is accepted to the IEEE Conference on Computer Vision and Pattern Recognition Workshop(CVPR

78 Dec 29, 2022

The 1st place solution of track2 (Vehicle Re-Identification) in the NVIDIA AI City Challenge at CVPR 2021 Workshop.

AICITY2021_Track2_DMT The 1st place solution of track2 (Vehicle Re-Identification) in the NVIDIA AI City Challenge at CVPR 2021 Workshop. Introduction

91 Dec 21, 2022

City-seeds - A random generator of cultural characteristics intended to spark ideas and help draw threads

City Seeds This is a random generator of cultural characteristics intended to sp

2 Mar 12, 2022

Code for "Multi-View Multi-Person 3D Pose Estimation with Plane Sweep Stereo"

Multi-View Multi-Person 3D Pose Estimation with Plane Sweep Stereo This repository includes the source code for our CVPR 2021 paper on multi-view mult

66 Jan 4, 2023

(CVPR 2022 - oral) Multi-View Depth Estimation by Fusing Single-View Depth Probability with Multi-View Geometry

Multi-View Depth Estimation by Fusing Single-View Depth Probability with Multi-View Geometry Official implementation of the paper Multi-View Depth Est

138 Dec 28, 2022

Web service for facial landmark detection, head pose estimation, facial action unit recognition, and eye-gaze estimation based on OpenFace 2.0

OpenGaze: Web Service for OpenFace Facial Behaviour Analysis Toolkit Overview OpenFace is a fantastic tool intended for computer vision and machine le

4 Nov 3, 2022

OpenPose: Real-time multi-person keypoint detection library for body, face, hands, and foot estimation

Build Type Linux MacOS Windows Build Status OpenPose has represented the first real-time multi-person system to jointly detect human body, hand, facia

25.7k Jan 9, 2023

MetaTTE: a Meta-Learning Based Travel Time Estimation Model for Multi-city Scenarios

Related tags

Overview

MetaTTE: a Meta-Learning Based Travel Time Estimation Model for Multi-city Scenarios

Core Requirements

Data Preparation

Configuration

Model Training

Eval baseline methods

Citation

You might also like...

CRISCE: Automatically Generating Critical Driving Scenarios From Car Accident Sketches

Time-series-deep-learning - Developing Deep learning LSTM, BiLSTM models, and NeuralProphet for multi-step time-series forecasting of stock price.

This repo is developed for Strong Baseline For Vehicle Re-Identification in Track 2 Ai-City-2021 Challenges

The 1st place solution of track2 (Vehicle Re-Identification) in the NVIDIA AI City Challenge at CVPR 2021 Workshop.

City-seeds - A random generator of cultural characteristics intended to spark ideas and help draw threads

Code for "Multi-View Multi-Person 3D Pose Estimation with Plane Sweep Stereo"

(CVPR 2022 - oral) Multi-View Depth Estimation by Fusing Single-View Depth Probability with Multi-View Geometry

Web service for facial landmark detection, head pose estimation, facial action unit recognition, and eye-gaze estimation based on OpenFace 2.0

OpenPose: Real-time multi-person keypoint detection library for body, face, hands, and foot estimation

Owner

morningstarwang

Re-implementation of the Noise Contrastive Estimation algorithm for pyTorch, following "Noise-contrastive estimation: A new estimation principle for unnormalized statistical models." (Gutmann and Hyvarinen, AISTATS 2010)

Out-of-Town Recommendation with Travel Intention Modeling (AAAI2021)

An open-source, low-cost, image-based weed detection device for fallow scenarios.

TPH-YOLOv5: Improved YOLOv5 Based on Transformer Prediction Head for Object Detection on Drone-Captured Scenarios

MetaDrive: Composing Diverse Scenarios for Generalizable Reinforcement Learning

Breaching - Breaching privacy in federated learning scenarios for vision and text

Code for the ICML 2021 paper "Bridging Multi-Task Learning and Meta-Learning: Towards Efficient Training and Effective Adaptation", Haoxiang Wang, Han Zhao, Bo Li.

Drone-based Joint Density Map Estimation, Localization and Tracking with Space-Time Multi-Scale Attention Network

Light-weight network, depth estimation, knowledge distillation, real-time depth estimation, auxiliary data.

PyTorch implementation of the supervised learning experiments from the paper Model-Agnostic Meta-Learning (MAML)