The model was trained with icefall (https://github.com/k2-fsa/icefall/pull/13), k2 (https://github.com/k2-fsa/k2/commit/81cec9ec736d2c603ad75d933bb3e3a3706fb0dd) and lhotse (https://github.com/lhotse-speech/lhotse/commit/5dfe0f4c02b1334ebb7db6d67e1141fe406ca76b)
See (https://github.com/k2-fsa/icefall/pull/13) for more details of this model.