Internet of Things (IoT) technology has enabled unprecedented smart applications made possible. However, it also brought increased security risks as the technology is the amalgamation of heterogeneous artefacts and lack of global standards yet. In this context, it is indispensable to have mechanisms to safeguard IoT applications from ever increasing cyberattacks. The emergence of artificial intelligence (AI) paved way for solving many real world problems. In this paper, we proposed a deep learning based framework known as Learning based Cyberattack Detection Framework (LbCADF) which exploits an enhanced Convolutional Neural Network (CNN) for automatic detection of cyberattacks in IoT use cases. The framework is capable of detecting attack traffic flows from benign ones. We proposed an algorithm known as Enhanced CNN for Attack Detection and Classification (ECNN-ADC). Our algorithm exploits feature selectin and hyperparameter tuning for leveraging quality of training. We configured early stopping criterion to get rid of overfitting. The proposed framework is evaluated using a benchmark dataset known as CICIDS2017. Our empirical study has revealed that the ECNN-ADC outperforms many state of the art models such as MLP and baseline CNN with highest accuracy 95% in cyberattack detection. The abstract should summarize the contents of the paper and should contain at least 70 and at most 150 words. It should be set in 9-point font size and should be inset 1.0 cm from the right and left margins. There should be two blank (10-point) lines before and after the abstract. This document is in the required format.