This initial corpus release is based on Coq 8.10.2 and Mathematical Components 1.9.0, and contains 449 source files from 21 Coq projects - in total over 297k lines of code (LOC). For each source file, there are two corresponding files with lists of S-expressions (sexps), organized at the Coq sentence level, for tokens and abstract syntax trees, respectively. Moreover, three sexp representations are provided for each Coq lemma statement in the corpus: tokens, abstract syntax tree, and elaborated term. All machine-readable representations were created using the SerAPI library version 0.7.1. The corpus release accompanies the research paper Deep Generation of Coq Lemma Names Using Elaborated Terms, published in the proceedings of IJCAR 2020.