Understanding macromolecular structures, such as proteins and nucleic acids, is critical for discerning their functions and biological roles. Advanced techniques like crystallography, NMR, and CryoEM have facilitated the determination of over 180,000 protein structures, all cataloged in the Protein Data Bank (PDB). This comprehensive repository has been pivotal in developing deep learning algorithms for predicting protein structures directly from sequences. In contrast, RNA structure prediction has lagged, primarily due to a scarcity of RNA structural data. Here, we present the secondary structures of 1098 pre-miRNAs and 1456 human mRNA regions determined through chemical probing. We develop a novel deep learning architecture, inspired from the Evoformer model of Alphafold and traditional architectures for secondary structure prediction. This new model, eFold, was trained on our newly generated database and over 100,000 secondary structures across multiple sources. We benchmark eFold on a set of challenging RNA structures and show that both our new architecture and dataset contributes to increasing the prediction performance compared to state of the art end-to-end methods. All together, these results reveal that merely expanding the database size is inadequate; rather, incorporating a greater diversity and complexity of structures is crucial for enhancing performance.