The files contain checkpoints of pruned and original text-to-audio generation model, AudioLDM-M-Full. The codes associated with the checkpoints are included here: Github Link. A description of attached files is below: audioldm-m-full.ckpt: pre-trained AudioLDM-M-Full checkpoint Unet_model-m.ckpt : pre-trained U-Net model checkpoint l1_audioldm-m-full_p1.ckpt: Pruned checkpoints, with b4 (or p) = 1, {b1,b2,b3} = {1,2,3} l1_audioldm-m-full_p1_dp1.ckpt: Pruned checkpoints, with b4 (or p) = 1, b3 (or dp) =1, {b1,b2} = {1,2} l1_audioldm-m-full_p2_dp2.ckpt: Pruned checkpoints, with b4 (or p) = 2, b3 (or dp) =2, {b1,b2} = {1,2} sorted_indexes_dict.pkl: sorted indxes across convolutional layers of U-Net using $l_1$-norm based pruning method.