Model Database's logo

sayakpaul
/
glpn-nyu-finetuned-diode-221228-072509

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

glpn-nyu-finetuned-diode-221228-072509

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: 0.4012
  • Mae: 0.4030
  • Rmse: 0.6173
  • Abs Rel: 0.3487
  • Log Mae: 0.1574
  • Log Rmse: 0.2110
  • Delta1: 0.4308
  • Delta2: 0.6997
  • Delta3: 0.8249

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: 0.0001
  • 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: 50
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Mae Rmse Abs Rel Log Mae Log Rmse Delta1 Delta2 Delta3
1.1571 1.0 72 0.6604 0.6233 0.8403 0.5125 0.3119 0.3691 0.1726 0.3423 0.4877
0.4895 2.0 144 0.4506 0.4460 0.6404 0.4241 0.1812 0.2299 0.3325 0.6053 0.7943
0.4709 3.0 216 0.4414 0.4370 0.6305 0.4243 0.1764 0.2253 0.3537 0.6145 0.7988
0.4436 4.0 288 0.4335 0.4324 0.6285 0.4045 0.1746 0.2245 0.3444 0.6506 0.8096
0.4656 5.0 360 0.4552 0.4515 0.6328 0.4614 0.1838 0.2307 0.3374 0.5762 0.7722
0.4482 6.0 432 0.4234 0.4166 0.6233 0.3805 0.1654 0.2179 0.4035 0.6623 0.8130
0.4099 7.0 504 0.4176 0.4185 0.6238 0.3676 0.1662 0.2150 0.3937 0.6589 0.8153
0.3987 8.0 576 0.4515 0.4431 0.6300 0.4497 0.1792 0.2283 0.3561 0.5906 0.7781
0.396 9.0 648 0.4235 0.4267 0.6347 0.3591 0.1716 0.2224 0.3934 0.6310 0.7963
0.3608 10.0 720 0.4312 0.4181 0.6227 0.4022 0.1666 0.2217 0.4014 0.6586 0.8173
0.3568 11.0 792 0.4322 0.4198 0.6183 0.4047 0.1674 0.2186 0.3870 0.6420 0.8071
0.3923 12.0 864 0.4225 0.4196 0.6294 0.3630 0.1668 0.2181 0.3910 0.6537 0.8151
0.3971 13.0 936 0.4086 0.4105 0.6219 0.3541 0.1614 0.2144 0.4234 0.6820 0.8144
0.372 14.0 1008 0.4127 0.4099 0.6172 0.3668 0.1612 0.2119 0.4046 0.6727 0.8260
0.3884 15.0 1080 0.4060 0.4074 0.6176 0.3528 0.1598 0.2119 0.4109 0.6925 0.8225
0.3616 16.0 1152 0.4078 0.4092 0.6198 0.3532 0.1615 0.2139 0.4162 0.6791 0.8186
0.3504 17.0 1224 0.4202 0.4320 0.6408 0.3613 0.1740 0.2261 0.3769 0.6301 0.7915
0.3823 18.0 1296 0.4328 0.4218 0.6182 0.4198 0.1684 0.2207 0.3916 0.6371 0.8113
0.3437 19.0 1368 0.4133 0.4138 0.6205 0.3638 0.1636 0.2162 0.3967 0.6761 0.8188
0.3739 20.0 1440 0.4040 0.4070 0.6187 0.3486 0.1594 0.2124 0.4214 0.6813 0.8214
0.3397 21.0 1512 0.4180 0.4300 0.6360 0.3601 0.1732 0.2239 0.3708 0.6362 0.8006
0.332 22.0 1584 0.4025 0.4050 0.6182 0.3505 0.1582 0.2114 0.4274 0.6909 0.8275
0.3552 23.0 1656 0.4120 0.4179 0.6305 0.3569 0.1650 0.2188 0.4002 0.6753 0.8102
0.3804 24.0 1728 0.4093 0.4111 0.6223 0.3594 0.1620 0.2152 0.4068 0.6851 0.8166
0.3519 25.0 1800 0.4039 0.4122 0.6237 0.3511 0.1621 0.2137 0.4109 0.6895 0.8171
0.3276 26.0 1872 0.4044 0.4117 0.6183 0.3533 0.1623 0.2127 0.3979 0.6824 0.8251
0.3167 27.0 1944 0.4091 0.4099 0.6189 0.3600 0.1613 0.2135 0.4069 0.6898 0.8218
0.3547 28.0 2016 0.4051 0.4055 0.6192 0.3521 0.1586 0.2119 0.4216 0.6921 0.8256
0.3297 29.0 2088 0.4025 0.4091 0.6215 0.3500 0.1605 0.2126 0.4155 0.6960 0.8224
0.3305 30.0 2160 0.4040 0.4045 0.6171 0.3507 0.1584 0.2120 0.4281 0.6938 0.8255
0.34 31.0 2232 0.4036 0.4082 0.6194 0.3492 0.1606 0.2132 0.4196 0.6851 0.8207
0.3507 32.0 2304 0.4057 0.4120 0.6245 0.3482 0.1619 0.2148 0.4195 0.6777 0.8172
0.3617 33.0 2376 0.4036 0.4098 0.6241 0.3477 0.1606 0.2141 0.4219 0.6871 0.8186
0.3268 34.0 2448 0.4015 0.4060 0.6197 0.3440 0.1593 0.2122 0.4326 0.6868 0.8211
0.3188 35.0 2520 0.4018 0.4032 0.6154 0.3504 0.1575 0.2107 0.4306 0.6952 0.8250
0.3286 36.0 2592 0.4046 0.4103 0.6237 0.3507 0.1611 0.2139 0.4179 0.6883 0.8173
0.3279 37.0 2664 0.3995 0.3993 0.6118 0.3460 0.1558 0.2091 0.4401 0.6979 0.8272
0.3439 38.0 2736 0.4052 0.4063 0.6196 0.3555 0.1590 0.2117 0.4207 0.6972 0.8256
0.3188 39.0 2808 0.4028 0.4028 0.6176 0.3482 0.1574 0.2112 0.4351 0.6916 0.8253
0.3334 40.0 2880 0.4059 0.4093 0.6218 0.3534 0.1607 0.2137 0.4201 0.6885 0.8217
0.3393 41.0 2952 0.4043 0.4048 0.6193 0.3492 0.1584 0.2118 0.4300 0.6906 0.8246
0.3099 42.0 3024 0.4029 0.4041 0.6161 0.3499 0.1583 0.2118 0.4274 0.6966 0.8239
0.3339 43.0 3096 0.4032 0.4056 0.6213 0.3515 0.1584 0.2122 0.4257 0.6995 0.8239
0.3086 44.0 3168 0.4024 0.4049 0.6173 0.3509 0.1586 0.2120 0.4243 0.6994 0.8227
0.3262 45.0 3240 0.4007 0.4035 0.6185 0.3467 0.1575 0.2112 0.4304 0.6994 0.8246
0.3265 46.0 3312 0.4017 0.4033 0.6170 0.3495 0.1574 0.2110 0.4271 0.7043 0.8247
0.3324 47.0 3384 0.4015 0.4056 0.6192 0.3471 0.1587 0.2119 0.4281 0.6944 0.8220
0.3159 48.0 3456 0.4012 0.4036 0.6156 0.3487 0.1581 0.2114 0.4279 0.6982 0.8234
0.3238 49.0 3528 0.4017 0.4024 0.6161 0.3499 0.1571 0.2106 0.4304 0.7008 0.8255
0.3112 50.0 3600 0.4012 0.4030 0.6173 0.3487 0.1574 0.2110 0.4308 0.6997 0.8249

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