Abstract: Hardware transient faults are proven to have a significant impact on deep neural networks (DNNs), whose safety-critical-misclassification probabilities in autonomous vehicles, healthcare, and space applications are increased up to 4x. However, the inaccuracy evaluation using accurate fault injection is time-consuming and requires several hours and even a couple of days on a complete simulation platform. To accelerate the evaluation of hardware transient faults on DNNs, we design a unified and end-to-end automatic methodology, A-Mean, to take advantage of the silent data corruption (SDC) rates of basic operations, such as convolution, add, multiply, Relu, Maxpooling, etc., and a two-level mean mechanism to rapidly compute the overall SDC rate for estimating the general classification metric, accuracy and application-specific metric safety-critical-misclassification (SCM). More importantly, a max policy is used to determine the SDC boundary of non-sequential structures in DNNs. Then, the worst-case scheme is also used to further calculate the enlarged SCM and halved accuracy under transient faults via merging the static results of SDC with the original data from one-time dynamic fault-free execution. Furthermore, all of the steps mentioned above have been implemented automatically so that this easy-to-use automatic tool can be employed for the prompt evaluation of transient faults on diverse DNNs. Meanwhile, a novel metric fault sensitivity is defined to jointly characterize the variation of transient fault-induced higher SCM and lower accuracy. The comparative results with a state-of-the-art fault injection method on five DNN models and four datasets show that our proposed estimation method A-Mean achieves up to 922.80x speedup, with just 4.20% SCM loss and 0.77% accuracy loss on average.