Runtime Neural Pruning
نویسندگان
چکیده
In this paper, we propose a Runtime Neural Pruning (RNP) framework which prunes the deep neural network dynamically at the runtime. Unlike existing neural pruning methods which produce a fixed pruned model for deployment, our method preserves the full ability of the original network and conducts pruning according to the input image and current feature maps adaptively. The pruning is performed in a bottom-up, layer-by-layer manner, which we model as a Markov decision process and use reinforcement learning for training. The agent judges the importance of each convolutional kernel and conducts channel-wise pruning conditioned on different samples, where the network is pruned more when the image is easier for the task. Since the ability of network is fully preserved, the balance point is easily adjustable according to the available resources. Our method can be applied to off-the-shelf network structures and reach a better tradeoff between speed and accuracy, especially with a large pruning rate.
منابع مشابه
Adaptive Predictive Controllers Using a Growing and Pruning RBF Neural Network
An adaptive version of growing and pruning RBF neural network has been used to predict the system output and implement Linear Model-Based Predictive Controller (LMPC) and Non-linear Model-based Predictive Controller (NMPC) strategies. A radial-basis neural network with growing and pruning capabilities is introduced to carry out on-line model identification.An Unscented Kal...
متن کاملExtracting Propositional Rules from Feed-forward Neural Networks - A New Decompositional Approach
In this paper, we present a new decompositional approach for the extraction of propositional rules from feed-forward neural networks of binary threshold units. After decomposing the network into single units, we show how to extract rules describing a unit’s behavior. This is done using a suitable search tree which allows the pruning of the search space. Furthermore, we present some experimental...
متن کاملBuild a Compact Binary Neural Network through Bit-level Sensitivity and Data Pruning
Convolutional neural network (CNN) has been widely used for vision-based tasks. Due to the high computational complexity and memory storage requirement, it is hard to directly deploy a full-precision CNN on embedded devices. The hardware-friendly designs are needed for re-source-limited and energy-constrained embed-ded devices. Emerging solutions are adopted for the neural network compression, ...
متن کاملAn Investigation into the performance and representations of a Stochastic, Evolutionary Neural Tree
The Stochastic Competitive Evolutionary Neural Tree (SCENT) is a new unsupervised neural net that dynamically evolves a representational structure in response to its training data. Uniquely SCENT requires no initial parameter setting as it autonomously creates appropriate parameterisation at runtime. Pruning and convergence are stochastically controlled using locally calculated heuristics. A th...
متن کاملSystem Identification with General Dynamic Neural Networks and Network Pruning
This paper presents an exact pruning algorithm with adaptive pruning interval for general dynamic neural networks (GDNN). GDNNs are artificial neural networks with internal dynamics. All layers have feedback connections with time delays to the same and to all other layers. The structure of the plant is unknown, so the identification process is started with a larger network architecture than nec...
متن کامل