glpn-nyu-finetuned-diode-221221-110911
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: 0.4188
- Mae: 0.4087
- Rmse: 0.6260
- Abs Rel: 0.3672
- Log Mae: 0.1626
- Log Rmse: 0.2222
- Delta1: 0.4391
- Delta2: 0.6801
- Delta3: 0.8037
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: 0.0005
- 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: 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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.5004 | 1.0 | 72 | 0.4287 | 0.4162 | 0.6314 | 0.3823 | 0.1665 | 0.2260 | 0.4303 | 0.6746 | 0.7946 |
| 0.3996 | 2.0 | 144 | 0.4796 | 0.4471 | 0.6517 | 0.4339 | 0.1868 | 0.2450 | 0.3620 | 0.5981 | 0.7637 |
| 0.4145 | 3.0 | 216 | 0.4401 | 0.4195 | 0.6352 | 0.4073 | 0.1682 | 0.2302 | 0.4202 | 0.6745 | 0.8016 |
| 0.4064 | 4.0 | 288 | 0.4573 | 0.4302 | 0.6366 | 0.4508 | 0.1742 | 0.2363 | 0.3917 | 0.6483 | 0.7958 |
| 0.4247 | 5.0 | 360 | 0.4441 | 0.4204 | 0.6227 | 0.4272 | 0.1693 | 0.2276 | 0.3986 | 0.6512 | 0.8034 |
| 0.4084 | 6.0 | 432 | 0.4326 | 0.4155 | 0.6287 | 0.3913 | 0.1667 | 0.2266 | 0.4183 | 0.6714 | 0.8022 |
| 0.3569 | 7.0 | 504 | 0.4334 | 0.4189 | 0.6355 | 0.3775 | 0.1697 | 0.2285 | 0.4125 | 0.6622 | 0.7967 |
| 0.3239 | 8.0 | 576 | 0.4246 | 0.4116 | 0.6259 | 0.3822 | 0.1640 | 0.2237 | 0.4319 | 0.6736 | 0.8043 |
| 0.32 | 9.0 | 648 | 0.4226 | 0.4171 | 0.6359 | 0.3648 | 0.1670 | 0.2256 | 0.4241 | 0.6677 | 0.7977 |
| 0.3087 | 10.0 | 720 | 0.4188 | 0.4087 | 0.6260 | 0.3672 | 0.1626 | 0.2222 | 0.4391 | 0.6801 | 0.8037 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.12.1+cu116
- Datasets 2.8.0
- Tokenizers 0.13.2
- Downloads last month
- 0