wandb: WARNING The get_url method is deprecated and will be removed in a future release. Please use `run.url` instead. [wandb] logging to https://wandb.ai/seungeun-goodgang-labs/animasync-v3-face/runs/v5c7fnv9 epoch 0 train l1=0.1973 vel=0.0200 val l1=0.0788 vel=0.0062 lr=3.74e-04 5.3s → saved best (val l1=0.0788) epoch 1 train l1=0.0807 vel=0.0077 val l1=0.0745 vel=0.0087 lr=7.48e-04 4.6s → saved best (val l1=0.0745) epoch 2 train l1=0.0806 vel=0.0077 val l1=0.0727 vel=0.0082 lr=1.00e-03 4.6s → saved best (val l1=0.0727) epoch 3 train l1=0.0772 vel=0.0081 val l1=0.0632 vel=0.0101 lr=9.98e-04 4.6s → saved best (val l1=0.0632) epoch 4 train l1=0.0624 vel=0.0106 val l1=0.0500 vel=0.0102 lr=9.94e-04 4.5s → saved best (val l1=0.0500) epoch 5 train l1=0.0460 vel=0.0121 val l1=0.0377 vel=0.0087 lr=9.88e-04 4.6s → saved best (val l1=0.0377) epoch 6 train l1=0.0385 vel=0.0106 val l1=0.0382 vel=0.0080 lr=9.80e-04 4.5s epoch 7 train l1=0.0344 vel=0.0095 val l1=0.0321 vel=0.0069 lr=9.69e-04 4.5s → saved best (val l1=0.0321) epoch 8 train l1=0.0330 vel=0.0090 val l1=0.0296 vel=0.0067 lr=9.57e-04 4.5s → saved best (val l1=0.0296) epoch 9 train l1=0.0307 vel=0.0086 val l1=0.0299 vel=0.0064 lr=9.42e-04 4.6s epoch 10 train l1=0.0295 vel=0.0083 val l1=0.0299 vel=0.0066 lr=9.26e-04 4.7s epoch 11 train l1=0.0285 vel=0.0081 val l1=0.0285 vel=0.0062 lr=9.07e-04 4.7s → saved best (val l1=0.0285) epoch 12 train l1=0.0234 vel=0.0081 val l1=0.0227 vel=0.0060 lr=8.87e-04 4.5s → saved best (val l1=0.0227) epoch 13 train l1=0.0215 vel=0.0077 val l1=0.0217 vel=0.0060 lr=8.65e-04 4.4s → saved best (val l1=0.0217) epoch 14 train l1=0.0213 vel=0.0077 val l1=0.0211 vel=0.0060 lr=8.42e-04 4.4s → saved best (val l1=0.0211) epoch 15 train l1=0.0206 vel=0.0076 val l1=0.0209 vel=0.0060 lr=8.17e-04 4.4s → saved best (val l1=0.0209) epoch 16 train l1=0.0200 vel=0.0074 val l1=0.0211 vel=0.0059 lr=7.90e-04 4.7s epoch 17 train l1=0.0197 vel=0.0074 val l1=0.0208 vel=0.0059 lr=7.63e-04 4.5s → saved best (val l1=0.0208) epoch 18 train l1=0.0192 vel=0.0073 val l1=0.0207 vel=0.0059 lr=7.34e-04 4.6s → saved best (val l1=0.0207) epoch 19 train l1=0.0188 vel=0.0073 val l1=0.0202 vel=0.0060 lr=7.04e-04 4.6s → saved best (val l1=0.0202) epoch 20 train l1=0.0186 vel=0.0072 val l1=0.0199 vel=0.0059 lr=6.73e-04 4.4s → saved best (val l1=0.0199) epoch 21 train l1=0.0182 vel=0.0072 val l1=0.0199 vel=0.0059 lr=6.42e-04 4.3s → saved best (val l1=0.0199) epoch 22 train l1=0.0179 vel=0.0071 val l1=0.0201 vel=0.0060 lr=6.10e-04 4.5s epoch 23 train l1=0.0176 vel=0.0071 val l1=0.0198 vel=0.0058 lr=5.77e-04 4.6s → saved best (val l1=0.0198) epoch 24 train l1=0.0174 vel=0.0070 val l1=0.0194 vel=0.0058 lr=5.44e-04 4.7s → saved best (val l1=0.0194) epoch 25 train l1=0.0172 vel=0.0070 val l1=0.0192 vel=0.0059 lr=5.11e-04 4.7s → saved best (val l1=0.0192) epoch 26 train l1=0.0169 vel=0.0070 val l1=0.0194 vel=0.0059 lr=4.78e-04 4.6s epoch 27 train l1=0.0168 vel=0.0070 val l1=0.0191 vel=0.0059 lr=4.45e-04 4.5s → saved best (val l1=0.0191) epoch 28 train l1=0.0165 vel=0.0069 val l1=0.0191 vel=0.0060 lr=4.12e-04 4.4s epoch 29 train l1=0.0164 vel=0.0069 val l1=0.0189 vel=0.0058 lr=3.80e-04 4.6s → saved best (val l1=0.0189) epoch 30 train l1=0.0162 vel=0.0069 val l1=0.0190 vel=0.0059 lr=3.48e-04 4.4s epoch 31 train l1=0.0160 vel=0.0069 val l1=0.0189 vel=0.0058 lr=3.16e-04 4.7s epoch 32 train l1=0.0158 vel=0.0068 val l1=0.0188 vel=0.0058 lr=2.86e-04 4.8s → saved best (val l1=0.0188) epoch 33 train l1=0.0156 vel=0.0068 val l1=0.0187 vel=0.0059 lr=2.56e-04 4.7s → saved best (val l1=0.0187) epoch 34 train l1=0.0156 vel=0.0068 val l1=0.0188 vel=0.0058 lr=2.28e-04 3.7s epoch 35 train l1=0.0154 vel=0.0068 val l1=0.0187 vel=0.0058 lr=2.01e-04 3.6s → saved best (val l1=0.0187) epoch 36 train l1=0.0153 vel=0.0068 val l1=0.0187 vel=0.0058 lr=1.75e-04 3.8s epoch 37 train l1=0.0152 vel=0.0068 val l1=0.0186 vel=0.0059 lr=1.50e-04 3.8s → saved best (val l1=0.0186) epoch 38 train l1=0.0150 vel=0.0067 val l1=0.0186 vel=0.0058 lr=1.27e-04 3.7s epoch 39 train l1=0.0150 vel=0.0067 val l1=0.0186 vel=0.0058 lr=1.06e-04 3.7s → saved best (val l1=0.0186) epoch 40 train l1=0.0149 vel=0.0067 val l1=0.0185 vel=0.0058 lr=8.66e-05 3.8s → saved best (val l1=0.0185) epoch 41 train l1=0.0149 vel=0.0067 val l1=0.0184 vel=0.0057 lr=6.89e-05 3.7s → saved best (val l1=0.0184) epoch 42 train l1=0.0148 vel=0.0067 val l1=0.0185 vel=0.0058 lr=5.30e-05 3.6s epoch 43 train l1=0.0147 vel=0.0067 val l1=0.0184 vel=0.0058 lr=3.91e-05 3.6s epoch 44 train l1=0.0147 vel=0.0067 val l1=0.0185 vel=0.0058 lr=2.73e-05 3.7s epoch 45 train l1=0.0146 vel=0.0067 val l1=0.0183 vel=0.0058 lr=1.75e-05 3.6s → saved best (val l1=0.0183) epoch 46 train l1=0.0146 vel=0.0067 val l1=0.0184 vel=0.0058 lr=9.88e-06 3.5s epoch 47 train l1=0.0146 vel=0.0067 val l1=0.0184 vel=0.0058 lr=4.40e-06 3.5s epoch 48 train l1=0.0145 vel=0.0067 val l1=0.0184 vel=0.0058 lr=1.10e-06 3.6s epoch 49 train l1=0.0146 vel=0.0067 val l1=0.0183 vel=0.0058 lr=0.00e+00 3.5s Done. best val l1: 0.0183 checkpoints: /dataset/kemix-engine/package/face/animasync-face-v3/models/v3_face/checkpoints