This paper provides a comprehensive review of deep learning approaches for time series classification (TSC). It evaluates the current state-of-the-art performance of deep learning algorithms for TSC by presenting an empirical study of the most recent deep neural network (DNN) architectures. The authors give an overview of the most successful deep learning applications in various time series domains under a unified taxonomy of DNNs for TSC. They also provide an open source deep learning framework for the TSC community, where they implemented each of the compared approaches and evaluated them on a univariate TSC benchmark (the UCR/UEA archive) and 12 multivariate time series datasets. By training 8,730 deep learning models on 97 time series datasets, they propose the most exhaustive study of DNNs for TSC to date. The paper discusses the success of deep learning in various classification tasks, which motivated the recent utilization of deep learning models for TSC. It highlights the effectiveness of deep convolutional neural networks (CNNs) in computer vision and their application to TSC. The authors also discuss the use of recurrent neural networks (RNNs) and echo state networks (ESNs) for TSC. They propose a unified taxonomy that regroups the recent applications of DNNs for TSC in various domains under two main categories: generative and discriminative models. They detail the architecture of nine end-to-end deep learning models designed specifically for TSC. They evaluate these models on the univariate UCR/UEA archive benchmark and 12 multivariate time series classification datasets. They provide the community with an open source deep learning framework for TSC in which they have implemented all nine approaches. They investigate the use of Class Activation Map (CAM) in order to reduce DNNs' black-box effect and explain the different decisions taken by various models. The paper also discusses the challenges of using deep learning for TSC, including the computational intensity of HIVE-COTE, the difficulty of interpreting decisions made by DNNs, and the need for efficient training methods. The authors propose a unified framework for deep learning models for TSC, which includes a common framework in Python, Keras, and Tensorflow to train the deep learning models on a cluster of more than 60 GPUs. They also discuss the use of regularization methods such as l2-norm weight decay and Dropout to reduce overfitting. The paper concludes with a discussion of the main contributions of the paper, including the explanation of how deep learning can be adapted to one dimensional time series data, the proposal of a unified taxonomy for DNNs for TSC, the detailed architecture of nine end-to-end deep learning models, the evaluation of these models on the UCR/UEA archive and 12 multivariate time series datasets, and the provision of an open source deep learning framework for TSC.This paper provides a comprehensive review of deep learning approaches for time series classification (TSC). It evaluates the current state-of-the-art performance of deep learning algorithms for TSC by presenting an empirical study of the most recent deep neural network (DNN) architectures. The authors give an overview of the most successful deep learning applications in various time series domains under a unified taxonomy of DNNs for TSC. They also provide an open source deep learning framework for the TSC community, where they implemented each of the compared approaches and evaluated them on a univariate TSC benchmark (the UCR/UEA archive) and 12 multivariate time series datasets. By training 8,730 deep learning models on 97 time series datasets, they propose the most exhaustive study of DNNs for TSC to date. The paper discusses the success of deep learning in various classification tasks, which motivated the recent utilization of deep learning models for TSC. It highlights the effectiveness of deep convolutional neural networks (CNNs) in computer vision and their application to TSC. The authors also discuss the use of recurrent neural networks (RNNs) and echo state networks (ESNs) for TSC. They propose a unified taxonomy that regroups the recent applications of DNNs for TSC in various domains under two main categories: generative and discriminative models. They detail the architecture of nine end-to-end deep learning models designed specifically for TSC. They evaluate these models on the univariate UCR/UEA archive benchmark and 12 multivariate time series classification datasets. They provide the community with an open source deep learning framework for TSC in which they have implemented all nine approaches. They investigate the use of Class Activation Map (CAM) in order to reduce DNNs' black-box effect and explain the different decisions taken by various models. The paper also discusses the challenges of using deep learning for TSC, including the computational intensity of HIVE-COTE, the difficulty of interpreting decisions made by DNNs, and the need for efficient training methods. The authors propose a unified framework for deep learning models for TSC, which includes a common framework in Python, Keras, and Tensorflow to train the deep learning models on a cluster of more than 60 GPUs. They also discuss the use of regularization methods such as l2-norm weight decay and Dropout to reduce overfitting. The paper concludes with a discussion of the main contributions of the paper, including the explanation of how deep learning can be adapted to one dimensional time series data, the proposal of a unified taxonomy for DNNs for TSC, the detailed architecture of nine end-to-end deep learning models, the evaluation of these models on the UCR/UEA archive and 12 multivariate time series datasets, and the provision of an open source deep learning framework for TSC.