This is the Libirspeech "dev-clean" dataset that can also be obtained by running:
curl https://www.openslr.org/resources/12/dev-clean.tar.gz --output dev-clean.tar.gz tar xf dev-clean.tar.gz