Use opensmile feature in neural network end to end?

[KuzmaNg]

Hi, I wanna know if I can use Opensmile to extract audio features to train neural models end to end, without saving the features as files. If there's a way, please let me know, sorry to bother. Thanks.