Model Database's logo

sayakpaul
/
glpn-nyu-finetuned-diode-221228-113625

Depth Estimation Transformers PyTorch glpn vision generated_from_trainer License: apache-2.0
Model card Files Community
Edit model card

glpn-nyu-finetuned-diode-221228-113625

This model is a fine-tuned version of vinvino02/glpn-nyu on the diode-subset dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3996
  • Mae: 0.4013
  • Rmse: 0.6161
  • Abs Rel: 0.3535
  • Log Mae: 0.1568
  • Log Rmse: 0.2121
  • Delta1: 0.4381
  • Delta2: 0.7025
  • Delta3: 0.8196

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 0.0003
  • train_batch_size: 24
  • eval_batch_size: 48
  • seed: 2022
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.15
  • num_epochs: 75
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Mae Rmse Abs Rel Log Mae Log Rmse Delta1 Delta2 Delta3
1.0075 1.0 72 0.4809 0.4610 0.6461 0.5165 0.1901 0.2446 0.3157 0.5632 0.8017
0.4692 2.0 144 0.4432 0.4491 0.6531 0.3950 0.1821 0.2318 0.3347 0.6198 0.7910
0.4635 3.0 216 0.4361 0.4278 0.6252 0.4165 0.1715 0.2230 0.3780 0.6285 0.8090
0.4364 4.0 288 0.4255 0.4200 0.6222 0.3930 0.1673 0.2198 0.3824 0.6639 0.8206
0.4632 5.0 360 0.4376 0.4267 0.6241 0.4144 0.1708 0.2235 0.3806 0.6337 0.8122
0.4703 6.0 432 0.4340 0.4315 0.6354 0.3799 0.1740 0.2262 0.3788 0.6275 0.7945
0.4136 7.0 504 0.4453 0.4291 0.6368 0.4144 0.1726 0.2306 0.3965 0.6458 0.7965
0.394 8.0 576 0.4620 0.4440 0.6297 0.4728 0.1808 0.2336 0.3606 0.5832 0.7826
0.4073 9.0 648 0.4485 0.4372 0.6244 0.4439 0.1769 0.2266 0.3511 0.6010 0.8002
0.3967 10.0 720 0.4523 0.4320 0.6250 0.4606 0.1750 0.2307 0.3676 0.6255 0.8146
0.3797 11.0 792 0.4413 0.4360 0.6332 0.4047 0.1769 0.2258 0.3426 0.6277 0.8025
0.439 12.0 864 0.4544 0.4365 0.6356 0.4215 0.1768 0.2299 0.3561 0.6282 0.8050
0.4666 13.0 936 0.4349 0.4278 0.6267 0.3893 0.1729 0.2227 0.3615 0.6375 0.8053
0.4071 14.0 1008 0.4337 0.4220 0.6235 0.3822 0.1692 0.2202 0.3909 0.6376 0.8044
0.4359 15.0 1080 0.4259 0.4193 0.6266 0.3855 0.1669 0.2217 0.4022 0.6601 0.8100
0.39 16.0 1152 0.4268 0.4075 0.6161 0.3981 0.1605 0.2184 0.4214 0.6838 0.8205
0.3654 17.0 1224 0.4503 0.4461 0.6615 0.3791 0.1840 0.2417 0.3783 0.6161 0.7636
0.4256 18.0 1296 0.4743 0.4529 0.6319 0.5162 0.1852 0.2398 0.3461 0.5736 0.7490
0.372 19.0 1368 0.4462 0.4326 0.6443 0.4068 0.1752 0.2331 0.3875 0.6410 0.7922
0.41 20.0 1440 0.4351 0.4500 0.6579 0.3735 0.1849 0.2365 0.3460 0.6021 0.7751
0.3683 21.0 1512 0.4060 0.4084 0.6177 0.3495 0.1605 0.2107 0.4168 0.6702 0.8235
0.36 22.0 1584 0.4447 0.4517 0.6667 0.3788 0.1852 0.2414 0.3676 0.6122 0.7572
0.4257 23.0 1656 0.4297 0.4141 0.6180 0.4066 0.1646 0.2201 0.4134 0.6586 0.8105
0.4344 24.0 1728 0.4545 0.4312 0.6237 0.4587 0.1742 0.2296 0.3769 0.6137 0.8008
0.4057 25.0 1800 0.4161 0.4099 0.6175 0.3744 0.1619 0.2144 0.4100 0.6701 0.8231
0.3569 26.0 1872 0.4199 0.4120 0.6181 0.3840 0.1634 0.2177 0.4039 0.6765 0.8165
0.3479 27.0 1944 0.4327 0.4180 0.6174 0.4138 0.1668 0.2205 0.3912 0.6481 0.8230
0.3732 28.0 2016 0.4426 0.4291 0.6236 0.4296 0.1715 0.2237 0.3866 0.6186 0.7911
0.3554 29.0 2088 0.4112 0.4073 0.6180 0.3598 0.1607 0.2146 0.4281 0.6800 0.8189
0.3679 30.0 2160 0.4139 0.4078 0.6190 0.3702 0.1609 0.2165 0.4249 0.6823 0.8110
0.3703 31.0 2232 0.4143 0.4097 0.6176 0.3730 0.1618 0.2156 0.4153 0.6782 0.8162
0.3605 32.0 2304 0.4179 0.4177 0.6303 0.3711 0.1654 0.2210 0.4062 0.6823 0.8022
0.3761 33.0 2376 0.4027 0.4070 0.6222 0.3441 0.1595 0.2127 0.4371 0.6834 0.8125
0.3352 34.0 2448 0.4077 0.4029 0.6134 0.3692 0.1581 0.2130 0.4322 0.6855 0.8273
0.336 35.0 2520 0.4212 0.4246 0.6328 0.3780 0.1696 0.2238 0.3844 0.6716 0.8005
0.3414 36.0 2592 0.4139 0.4132 0.6241 0.3720 0.1639 0.2184 0.4162 0.6714 0.8092
0.3416 37.0 2664 0.4183 0.4101 0.6149 0.3844 0.1625 0.2159 0.4157 0.6649 0.8172
0.3765 38.0 2736 0.4207 0.4120 0.6199 0.3926 0.1635 0.2193 0.4066 0.6767 0.8154
0.3548 39.0 2808 0.4096 0.4056 0.6167 0.3667 0.1593 0.2138 0.4244 0.6905 0.8213
0.3822 40.0 2880 0.4084 0.4061 0.6180 0.3653 0.1593 0.2134 0.4246 0.6891 0.8249
0.3505 41.0 2952 0.4041 0.4118 0.6271 0.3515 0.1620 0.2156 0.4279 0.6872 0.8098
0.3514 42.0 3024 0.4033 0.4006 0.6185 0.3558 0.1563 0.2132 0.4510 0.7030 0.8181
0.3459 43.0 3096 0.4061 0.4051 0.6196 0.3631 0.1587 0.2147 0.4282 0.7019 0.8206
0.3213 44.0 3168 0.4041 0.4093 0.6232 0.3539 0.1605 0.2148 0.4301 0.6893 0.8168
0.3346 45.0 3240 0.4103 0.4023 0.6151 0.3705 0.1578 0.2141 0.4339 0.6907 0.8219
0.3585 46.0 3312 0.4054 0.3953 0.6096 0.3627 0.1542 0.2113 0.4524 0.7052 0.8251
0.3799 47.0 3384 0.4063 0.4100 0.6230 0.3574 0.1616 0.2165 0.4263 0.6821 0.8113
0.3235 48.0 3456 0.4051 0.4004 0.6117 0.3692 0.1571 0.2123 0.4364 0.6928 0.8268
0.3628 49.0 3528 0.4051 0.3985 0.6115 0.3622 0.1560 0.2111 0.4486 0.6932 0.8234
0.3399 50.0 3600 0.4145 0.4059 0.6184 0.3789 0.1598 0.2169 0.4260 0.6977 0.8194
0.3288 51.0 3672 0.4089 0.4057 0.6172 0.3692 0.1597 0.2153 0.4300 0.6939 0.8198
0.3231 52.0 3744 0.4104 0.4126 0.6261 0.3643 0.1628 0.2185 0.4296 0.6826 0.8104
0.3238 53.0 3816 0.4107 0.4023 0.6170 0.3745 0.1580 0.2167 0.4362 0.7031 0.8216
0.3253 54.0 3888 0.4056 0.4006 0.6135 0.3673 0.1570 0.2134 0.4400 0.7034 0.8221
0.3383 55.0 3960 0.4053 0.4060 0.6187 0.3598 0.1593 0.2141 0.4310 0.6938 0.8187
0.3279 56.0 4032 0.4118 0.4003 0.6130 0.3797 0.1569 0.2153 0.4388 0.7040 0.8212
0.32 57.0 4104 0.4042 0.4001 0.6185 0.3566 0.1560 0.2123 0.4470 0.7070 0.8227
0.3282 58.0 4176 0.4035 0.4010 0.6173 0.3533 0.1568 0.2126 0.4438 0.7037 0.8208
0.3271 59.0 4248 0.4015 0.4018 0.6168 0.3551 0.1570 0.2123 0.4334 0.7095 0.8201
0.3127 60.0 4320 0.4029 0.3975 0.6142 0.3590 0.1549 0.2113 0.4420 0.7082 0.8245
0.3142 61.0 4392 0.4044 0.4031 0.6163 0.3585 0.1577 0.2126 0.4273 0.7034 0.8214
0.3059 62.0 4464 0.4034 0.4033 0.6151 0.3624 0.1580 0.2127 0.4256 0.7038 0.8223
0.3133 63.0 4536 0.4028 0.4066 0.6205 0.3554 0.1594 0.2137 0.4235 0.6991 0.8187
0.3086 64.0 4608 0.4023 0.3982 0.6117 0.3588 0.1556 0.2108 0.4381 0.7002 0.8248
0.3143 65.0 4680 0.4036 0.4084 0.6250 0.3566 0.1600 0.2157 0.4323 0.6946 0.8094
0.3031 66.0 4752 0.4012 0.3999 0.6170 0.3551 0.1559 0.2122 0.4458 0.7044 0.8200
0.3279 67.0 4824 0.4031 0.4001 0.6160 0.3609 0.1562 0.2129 0.4421 0.7042 0.8205
0.3173 68.0 4896 0.4000 0.3989 0.6141 0.3569 0.1557 0.2120 0.4456 0.7040 0.8226
0.3203 69.0 4968 0.3989 0.3995 0.6153 0.3545 0.1556 0.2114 0.4421 0.7069 0.8215
0.3165 70.0 5040 0.3984 0.3993 0.6144 0.3513 0.1558 0.2111 0.4450 0.7027 0.8222
0.3278 71.0 5112 0.3993 0.4032 0.6191 0.3509 0.1574 0.2124 0.4386 0.7007 0.8184
0.3232 72.0 5184 0.3990 0.4000 0.6149 0.3534 0.1561 0.2112 0.4396 0.7018 0.8223
0.3089 73.0 5256 0.3996 0.4022 0.6172 0.3526 0.1571 0.2121 0.4370 0.7011 0.8197
0.3118 74.0 5328 0.3994 0.4016 0.6164 0.3530 0.1570 0.2121 0.4375 0.7026 0.8195
0.3161 75.0 5400 0.3996 0.4013 0.6161 0.3535 0.1568 0.2121 0.4381 0.7025 0.8196

Framework versions

  • Transformers 4.24.0
  • Pytorch 1.12.1+cu116
  • Datasets 2.8.0
  • Tokenizers 0.13.2
Downloads last month
0

Evaluation results

Metadata error: specify a dataset to view leaderboard