glpn-nyu-finetuned-diode-221116-104421
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: 0.3736
- Mae: 0.3079
- Rmse: 0.4321
- Abs Rel: 0.3666
- Log Mae: 0.1288
- Log Rmse: 0.1794
- Delta1: 0.4929
- Delta2: 0.7934
- Delta3: 0.9234
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: 1e-05
- 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: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Mae | Rmse | Abs Rel | Log Mae | Log Rmse | Delta1 | Delta2 | Delta3 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1.3644 | 1.0 | 72 | 1.1149 | 2.9299 | 2.9878 | 4.3237 | 0.6329 | 0.6803 | 0.0001 | 0.0241 | 0.1147 |
| 0.7701 | 2.0 | 144 | 0.5115 | 0.4977 | 0.6435 | 0.6806 | 0.1967 | 0.2543 | 0.3113 | 0.5462 | 0.7732 |
| 0.4351 | 3.0 | 216 | 0.4129 | 0.3591 | 0.4868 | 0.4587 | 0.1495 | 0.2034 | 0.4214 | 0.7194 | 0.8869 |
| 0.4001 | 4.0 | 288 | 0.4003 | 0.3421 | 0.4711 | 0.4248 | 0.1418 | 0.1953 | 0.4509 | 0.7446 | 0.8999 |
| 0.3923 | 5.0 | 360 | 0.3928 | 0.3334 | 0.4573 | 0.4139 | 0.1388 | 0.1906 | 0.4562 | 0.7570 | 0.9098 |
| 0.363 | 6.0 | 432 | 0.3806 | 0.3176 | 0.4419 | 0.3873 | 0.1328 | 0.1840 | 0.4757 | 0.7786 | 0.9188 |
| 0.3516 | 7.0 | 504 | 0.3760 | 0.3091 | 0.4346 | 0.3697 | 0.1291 | 0.1804 | 0.4933 | 0.7927 | 0.9224 |
| 0.303 | 8.0 | 576 | 0.3798 | 0.3131 | 0.4401 | 0.3811 | 0.1307 | 0.1833 | 0.4913 | 0.7886 | 0.9189 |
| 0.3191 | 9.0 | 648 | 0.3766 | 0.3104 | 0.4356 | 0.3738 | 0.1298 | 0.1811 | 0.4907 | 0.7901 | 0.9214 |
| 0.3102 | 10.0 | 720 | 0.3736 | 0.3079 | 0.4321 | 0.3666 | 0.1288 | 0.1794 | 0.4929 | 0.7934 | 0.9234 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.12.1+cu113
- Tokenizers 0.13.2
- Downloads last month
- 0