In the context of Deep Learning training, memory needs to store activations can prevent the user to consider large models and large batch sizes. A possible solution is to rely on model parallelism to distribute the weights of the model and the activations over distributedmemory nodes. In this paper, we consider another purely sequential approach to save memory using checkpointing techniques. Checkpointing techniques have been introduced in the context of Automatic Differentiation. They consist in storing some, but not all activations during the feed-forward network training phase, and then to recompute missing values during the backward phase. Using this approach, it is possible, at the price of ...
Deep learning methods have recently started dominating the machine learning world as they offer stat...
We propose a new integrated method of exploiting model, batch and domain parallelism for the trainin...
Residual connections are ubiquitous in deep learning, since besides residual networks and their vari...
In the context of Deep Learning training, memory needs to store activations can prevent ...
This paper introduces a new activation checkpointing method which allows to significantly decrease m...
International audienceDeep Learning training memory needs can preventthe user to consider large mode...
Artificial Intelligence is a field that has received a lot of attention recently. Its success is due...
The training phase in Deep Neural Networks has become an important source of computing resource usag...
International audienceRematerialization and offloading are two well known strategies to save memory ...
International audienceTraining Deep Neural Networks is known to be an expensive operation, both in t...
We are interested in the problem of continual learning of artificial neural networks in the case whe...
Scientific workflows are frequently modeled as Directed Acyclic Graphs (DAG) oftasks, which repres...
International audienceWith the emergence of versatile storage systems, multi-level checkpointing (ML...
To make machine learning (ML) sustainable and apt to run on the diverse devices where relevant data...
The compilation of high-level programming languages for parallel machines faces two challenges: maxi...
Deep learning methods have recently started dominating the machine learning world as they offer stat...
We propose a new integrated method of exploiting model, batch and domain parallelism for the trainin...
Residual connections are ubiquitous in deep learning, since besides residual networks and their vari...
In the context of Deep Learning training, memory needs to store activations can prevent ...
This paper introduces a new activation checkpointing method which allows to significantly decrease m...
International audienceDeep Learning training memory needs can preventthe user to consider large mode...
Artificial Intelligence is a field that has received a lot of attention recently. Its success is due...
The training phase in Deep Neural Networks has become an important source of computing resource usag...
International audienceRematerialization and offloading are two well known strategies to save memory ...
International audienceTraining Deep Neural Networks is known to be an expensive operation, both in t...
We are interested in the problem of continual learning of artificial neural networks in the case whe...
Scientific workflows are frequently modeled as Directed Acyclic Graphs (DAG) oftasks, which repres...
International audienceWith the emergence of versatile storage systems, multi-level checkpointing (ML...
To make machine learning (ML) sustainable and apt to run on the diverse devices where relevant data...
The compilation of high-level programming languages for parallel machines faces two challenges: maxi...
Deep learning methods have recently started dominating the machine learning world as they offer stat...
We propose a new integrated method of exploiting model, batch and domain parallelism for the trainin...
Residual connections are ubiquitous in deep learning, since besides residual networks and their vari...