detr
This model is a fine-tuned version of facebook/detr-resnet-50 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6906
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: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
2.6507 | 0.04 | 100 | 2.1677 |
1.8892 | 0.08 | 200 | 1.6329 |
1.638 | 0.12 | 300 | 1.5517 |
1.4457 | 0.16 | 400 | 1.4176 |
1.3489 | 0.2 | 500 | 1.2559 |
1.277 | 0.24 | 600 | 1.2726 |
1.2948 | 0.28 | 700 | 1.2169 |
1.1878 | 0.32 | 800 | 1.1680 |
1.1781 | 0.36 | 900 | 1.1291 |
1.1747 | 0.4 | 1000 | 1.1146 |
1.1966 | 0.44 | 1100 | 1.1399 |
1.1641 | 0.48 | 1200 | 1.0844 |
1.128 | 0.52 | 1300 | 1.1119 |
1.1191 | 0.56 | 1400 | 1.0528 |
1.1435 | 0.6 | 1500 | 1.0689 |
1.1657 | 0.64 | 1600 | 1.1484 |
1.1727 | 0.68 | 1700 | 1.0764 |
1.1085 | 0.72 | 1800 | 1.0391 |
1.0579 | 0.76 | 1900 | 1.0012 |
1.0935 | 0.8 | 2000 | 1.0160 |
1.054 | 0.84 | 2100 | 0.9796 |
1.0486 | 0.88 | 2200 | 0.9508 |
1.0472 | 0.92 | 2300 | 0.9869 |
1.032 | 0.96 | 2400 | 0.9655 |
1.0313 | 1.0 | 2500 | 0.9484 |
1.0096 | 1.04 | 2600 | 0.9676 |
1.0279 | 1.08 | 2700 | 1.0771 |
1.027 | 1.12 | 2800 | 0.9912 |
1.0415 | 1.16 | 2900 | 0.9882 |
1.003 | 1.2 | 3000 | 0.9579 |
1.0084 | 1.24 | 3100 | 0.9288 |
0.9353 | 1.28 | 3200 | 0.9278 |
0.9514 | 1.32 | 3300 | 0.8915 |
0.9452 | 1.36 | 3400 | 0.8904 |
0.9312 | 1.4 | 3500 | 0.8925 |
0.9256 | 1.44 | 3600 | 0.8729 |
0.8861 | 1.48 | 3700 | 0.8655 |
0.9043 | 1.52 | 3800 | 0.8977 |
0.8935 | 1.56 | 3900 | 0.8679 |
0.8974 | 1.6 | 4000 | 0.8908 |
0.9342 | 1.64 | 4100 | 0.8742 |
0.889 | 1.68 | 4200 | 0.8534 |
0.8998 | 1.72 | 4300 | 0.8409 |
0.8727 | 1.76 | 4400 | 0.8333 |
0.8728 | 1.8 | 4500 | 0.8386 |
0.8525 | 1.84 | 4600 | 0.8152 |
0.8709 | 1.88 | 4700 | 0.8146 |
0.8694 | 1.92 | 4800 | 0.8245 |
0.8663 | 1.96 | 4900 | 0.8216 |
0.8442 | 2.0 | 5000 | 0.8019 |
0.8256 | 2.04 | 5100 | 0.8022 |
0.8385 | 2.08 | 5200 | 0.7938 |
0.7995 | 2.12 | 5300 | 0.7958 |
0.8217 | 2.16 | 5400 | 0.7962 |
0.8432 | 2.2 | 5500 | 0.7772 |
0.8228 | 2.24 | 5600 | 0.7857 |
0.8283 | 2.28 | 5700 | 0.7982 |
0.772 | 2.32 | 5800 | 0.7969 |
0.8019 | 2.36 | 5900 | 0.7902 |
0.7805 | 2.4 | 6000 | 0.7782 |
0.802 | 2.44 | 6100 | 0.7681 |
0.8483 | 2.48 | 6200 | 0.7722 |
0.802 | 2.52 | 6300 | 0.7673 |
0.8064 | 2.56 | 6400 | 0.7603 |
0.7638 | 2.6 | 6500 | 0.7475 |
0.7727 | 2.64 | 6600 | 0.7515 |
0.801 | 2.68 | 6700 | 0.7523 |
0.8022 | 2.72 | 6800 | 0.7519 |
0.8074 | 2.76 | 6900 | 0.7555 |
0.7951 | 2.8 | 7000 | 0.7450 |
0.8125 | 2.84 | 7100 | 0.7476 |
0.8085 | 2.88 | 7200 | 0.7505 |
0.7959 | 2.92 | 7300 | 0.7432 |
0.7668 | 2.96 | 7400 | 0.7454 |
0.7666 | 3.0 | 7500 | 0.7419 |
0.7422 | 3.04 | 7600 | 0.7284 |
0.7713 | 3.08 | 7700 | 0.7418 |
0.7296 | 3.12 | 7800 | 0.7274 |
0.7468 | 3.16 | 7900 | 0.7224 |
0.7767 | 3.2 | 8000 | 0.7268 |
0.7526 | 3.24 | 8100 | 0.7210 |
0.7328 | 3.28 | 8200 | 0.7139 |
0.7626 | 3.32 | 8300 | 0.7142 |
0.7515 | 3.36 | 8400 | 0.7102 |
0.7141 | 3.4 | 8500 | 0.7100 |
0.7068 | 3.44 | 8600 | 0.7097 |
0.7274 | 3.48 | 8700 | 0.7018 |
0.7458 | 3.52 | 8800 | 0.7041 |
0.7205 | 3.56 | 8900 | 0.7065 |
0.7643 | 3.6 | 9000 | 0.6985 |
0.6968 | 3.64 | 9100 | 0.6983 |
0.7111 | 3.68 | 9200 | 0.6982 |
0.7229 | 3.72 | 9300 | 0.6920 |
0.7466 | 3.76 | 9400 | 0.6959 |
0.7126 | 3.8 | 9500 | 0.6925 |
0.739 | 3.84 | 9600 | 0.6869 |
0.7449 | 3.88 | 9700 | 0.6939 |
0.7139 | 3.92 | 9800 | 0.6893 |
0.7216 | 3.96 | 9900 | 0.6895 |
0.6942 | 4.0 | 10000 | 0.6906 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2
- Downloads last month
- 39