<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
segformer-b0_DsMetalDam_NoBright_Augmented_Cropped
This model is a fine-tuned version of nvidia/mit-b0 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2340
- Mean Iou: 0.7058
- Mean Accuracy: 0.7765
- Overall Accuracy: 0.9189
- Accuracy Matrix: 0.9030
- Accuracy Austenite: 0.9494
- Accuracy Martensite/austenite: 0.8141
- Accuracy Precipitate: 0.2512
- Accuracy Defect: 0.9650
- Iou Matrix: 0.8221
- Iou Austenite: 0.8966
- Iou Martensite/austenite: 0.7116
- Iou Precipitate: 0.2073
- Iou Defect: 0.8915
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: 6e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
Training results
Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Accuracy Matrix | Accuracy Austenite | Accuracy Martensite/austenite | Accuracy Precipitate | Accuracy Defect | Iou Matrix | Iou Austenite | Iou Martensite/austenite | Iou Precipitate | Iou Defect |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0.4584 | 1.0 | 272 | 0.4505 | 0.5157 | 0.5969 | 0.8464 | 0.8792 | 0.9262 | 0.2774 | 0.0 | 0.9015 | 0.7053 | 0.8328 | 0.2627 | 0.0 | 0.7779 |
0.2925 | 2.0 | 544 | 0.3665 | 0.5763 | 0.6670 | 0.8655 | 0.8262 | 0.9330 | 0.6344 | 0.0 | 0.9413 | 0.7206 | 0.8501 | 0.4845 | 0.0 | 0.8266 |
0.3749 | 3.0 | 816 | 0.3568 | 0.5810 | 0.6814 | 0.8653 | 0.8274 | 0.9165 | 0.7248 | 0.0 | 0.9384 | 0.7212 | 0.8540 | 0.4993 | 0.0 | 0.8308 |
0.2736 | 4.0 | 1088 | 0.3249 | 0.5920 | 0.6726 | 0.8764 | 0.8387 | 0.9439 | 0.6405 | 0.0 | 0.9397 | 0.7400 | 0.8596 | 0.5157 | 0.0 | 0.8446 |
0.3164 | 5.0 | 1360 | 0.3070 | 0.6039 | 0.6895 | 0.8841 | 0.8875 | 0.9180 | 0.6928 | 0.0 | 0.9493 | 0.7610 | 0.8646 | 0.5548 | 0.0 | 0.8394 |
0.3564 | 6.0 | 1632 | 0.2998 | 0.6040 | 0.6990 | 0.8854 | 0.8503 | 0.9336 | 0.7525 | 0.0 | 0.9588 | 0.7572 | 0.8679 | 0.5741 | 0.0 | 0.8206 |
0.2259 | 7.0 | 1904 | 0.2936 | 0.6140 | 0.6860 | 0.8906 | 0.8955 | 0.9280 | 0.6752 | 0.0 | 0.9311 | 0.7696 | 0.8730 | 0.5698 | 0.0 | 0.8576 |
0.2193 | 8.0 | 2176 | 0.2777 | 0.6192 | 0.7030 | 0.8950 | 0.8824 | 0.9338 | 0.7338 | 0.0 | 0.9651 | 0.7763 | 0.8745 | 0.6090 | 0.0 | 0.8361 |
0.2692 | 9.0 | 2448 | 0.2774 | 0.6212 | 0.7061 | 0.8939 | 0.8731 | 0.9321 | 0.7680 | 0.0 | 0.9572 | 0.7734 | 0.8759 | 0.6053 | 0.0 | 0.8516 |
0.258 | 10.0 | 2720 | 0.2693 | 0.6260 | 0.7114 | 0.8974 | 0.8671 | 0.9387 | 0.7888 | 0.0005 | 0.9619 | 0.7787 | 0.8770 | 0.6305 | 0.0005 | 0.8433 |
0.2413 | 11.0 | 2992 | 0.2647 | 0.6317 | 0.7055 | 0.8998 | 0.8783 | 0.9420 | 0.7535 | 0.0010 | 0.9525 | 0.7835 | 0.8789 | 0.6321 | 0.0010 | 0.8632 |
0.2519 | 12.0 | 3264 | 0.2610 | 0.6362 | 0.7122 | 0.9019 | 0.8783 | 0.9426 | 0.7724 | 0.0070 | 0.9609 | 0.7877 | 0.8797 | 0.6479 | 0.0070 | 0.8586 |
0.2985 | 13.0 | 3536 | 0.2595 | 0.6508 | 0.7216 | 0.9021 | 0.8783 | 0.9431 | 0.7701 | 0.0642 | 0.9521 | 0.7881 | 0.8810 | 0.6419 | 0.0609 | 0.8823 |
0.1429 | 14.0 | 3808 | 0.2553 | 0.6552 | 0.7304 | 0.9045 | 0.8912 | 0.9321 | 0.8125 | 0.0623 | 0.9540 | 0.7948 | 0.8820 | 0.6609 | 0.0590 | 0.8794 |
0.3317 | 15.0 | 4080 | 0.2617 | 0.6484 | 0.7187 | 0.9048 | 0.8865 | 0.9475 | 0.7424 | 0.0527 | 0.9644 | 0.7945 | 0.8828 | 0.6469 | 0.0502 | 0.8678 |
0.2437 | 16.0 | 4352 | 0.2495 | 0.6572 | 0.7296 | 0.9071 | 0.8892 | 0.9383 | 0.8119 | 0.0545 | 0.9541 | 0.7994 | 0.8846 | 0.6702 | 0.0524 | 0.8794 |
0.2262 | 17.0 | 4624 | 0.2508 | 0.6634 | 0.7270 | 0.9081 | 0.8991 | 0.9415 | 0.7665 | 0.0911 | 0.9370 | 0.8016 | 0.8860 | 0.6627 | 0.0849 | 0.8819 |
0.3598 | 18.0 | 4896 | 0.2472 | 0.6695 | 0.7443 | 0.9072 | 0.8726 | 0.9468 | 0.8215 | 0.1279 | 0.9527 | 0.7965 | 0.8851 | 0.6747 | 0.1146 | 0.8769 |
0.2139 | 19.0 | 5168 | 0.2433 | 0.6613 | 0.7415 | 0.9083 | 0.8841 | 0.9401 | 0.8305 | 0.0859 | 0.9669 | 0.8004 | 0.8874 | 0.6747 | 0.0808 | 0.8634 |
0.2296 | 20.0 | 5440 | 0.2446 | 0.6838 | 0.7598 | 0.9096 | 0.8914 | 0.9428 | 0.7992 | 0.2142 | 0.9516 | 0.8037 | 0.8875 | 0.6764 | 0.1722 | 0.8795 |
0.1481 | 21.0 | 5712 | 0.2408 | 0.6820 | 0.7550 | 0.9113 | 0.8928 | 0.9423 | 0.8155 | 0.1627 | 0.9615 | 0.8064 | 0.8892 | 0.6859 | 0.1408 | 0.8876 |
0.1734 | 22.0 | 5984 | 0.2478 | 0.6653 | 0.7292 | 0.9114 | 0.9144 | 0.9353 | 0.7816 | 0.0636 | 0.9512 | 0.8092 | 0.8875 | 0.6850 | 0.0616 | 0.8835 |
0.197 | 23.0 | 6256 | 0.2379 | 0.6747 | 0.7441 | 0.9130 | 0.8951 | 0.9474 | 0.7965 | 0.1149 | 0.9668 | 0.8097 | 0.8906 | 0.6902 | 0.1068 | 0.8763 |
0.1571 | 24.0 | 6528 | 0.2483 | 0.6773 | 0.7489 | 0.9120 | 0.9064 | 0.9372 | 0.8033 | 0.1306 | 0.9672 | 0.8103 | 0.8889 | 0.6864 | 0.1184 | 0.8823 |
0.1778 | 25.0 | 6800 | 0.2405 | 0.6899 | 0.7690 | 0.9122 | 0.9057 | 0.9355 | 0.8155 | 0.2224 | 0.9658 | 0.8098 | 0.8895 | 0.6906 | 0.1818 | 0.8780 |
0.2093 | 26.0 | 7072 | 0.2403 | 0.6855 | 0.7525 | 0.9138 | 0.9001 | 0.9447 | 0.8018 | 0.1600 | 0.9558 | 0.8122 | 0.8908 | 0.6936 | 0.1415 | 0.8894 |
0.2055 | 27.0 | 7344 | 0.2384 | 0.6823 | 0.7530 | 0.9141 | 0.8960 | 0.9486 | 0.7962 | 0.1546 | 0.9698 | 0.8119 | 0.8919 | 0.6928 | 0.1384 | 0.8765 |
0.1709 | 28.0 | 7616 | 0.2367 | 0.6906 | 0.7603 | 0.9152 | 0.8969 | 0.9485 | 0.8045 | 0.1884 | 0.9630 | 0.8144 | 0.8924 | 0.6997 | 0.1633 | 0.8834 |
0.206 | 29.0 | 7888 | 0.2374 | 0.6922 | 0.7640 | 0.9150 | 0.9018 | 0.9492 | 0.7797 | 0.2197 | 0.9698 | 0.8145 | 0.8925 | 0.6937 | 0.1836 | 0.8765 |
0.3099 | 30.0 | 8160 | 0.2370 | 0.6978 | 0.7718 | 0.9155 | 0.8973 | 0.9448 | 0.8277 | 0.2267 | 0.9626 | 0.8150 | 0.8931 | 0.7038 | 0.1880 | 0.8894 |
0.2932 | 31.0 | 8432 | 0.2345 | 0.6972 | 0.7716 | 0.9159 | 0.9064 | 0.9419 | 0.8144 | 0.2279 | 0.9675 | 0.8171 | 0.8932 | 0.7023 | 0.1876 | 0.8859 |
0.1847 | 32.0 | 8704 | 0.2340 | 0.6942 | 0.7670 | 0.9155 | 0.8924 | 0.9484 | 0.8246 | 0.2059 | 0.9634 | 0.8146 | 0.8934 | 0.7017 | 0.1753 | 0.8860 |
0.2105 | 33.0 | 8976 | 0.2316 | 0.7034 | 0.7817 | 0.9162 | 0.9019 | 0.9451 | 0.8149 | 0.2827 | 0.9641 | 0.8174 | 0.8938 | 0.7024 | 0.2173 | 0.8862 |
0.1999 | 34.0 | 9248 | 0.2339 | 0.6958 | 0.7642 | 0.9170 | 0.9014 | 0.9491 | 0.8038 | 0.2001 | 0.9664 | 0.8181 | 0.8945 | 0.7054 | 0.1717 | 0.8895 |
0.137 | 35.0 | 9520 | 0.2334 | 0.7043 | 0.7798 | 0.9166 | 0.9006 | 0.9439 | 0.8308 | 0.2675 | 0.9560 | 0.8177 | 0.8942 | 0.7069 | 0.2120 | 0.8909 |
0.1199 | 36.0 | 9792 | 0.2339 | 0.7001 | 0.7739 | 0.9172 | 0.8972 | 0.9493 | 0.8187 | 0.2339 | 0.9704 | 0.8182 | 0.8952 | 0.7070 | 0.1958 | 0.8846 |
0.1345 | 37.0 | 10064 | 0.2346 | 0.7074 | 0.7820 | 0.9173 | 0.9003 | 0.9478 | 0.8158 | 0.2853 | 0.9607 | 0.8191 | 0.8949 | 0.7067 | 0.2227 | 0.8938 |
0.147 | 38.0 | 10336 | 0.2354 | 0.7011 | 0.7779 | 0.9169 | 0.8983 | 0.9457 | 0.8321 | 0.2449 | 0.9683 | 0.8181 | 0.8949 | 0.7070 | 0.2020 | 0.8835 |
0.1549 | 39.0 | 10608 | 0.2345 | 0.6981 | 0.7646 | 0.9176 | 0.9029 | 0.9505 | 0.7960 | 0.2098 | 0.9641 | 0.8197 | 0.8949 | 0.7054 | 0.1807 | 0.8897 |
0.2041 | 40.0 | 10880 | 0.2347 | 0.7018 | 0.7738 | 0.9176 | 0.8956 | 0.9515 | 0.8163 | 0.2390 | 0.9668 | 0.8187 | 0.8954 | 0.7086 | 0.1983 | 0.8878 |
0.1228 | 41.0 | 11152 | 0.2342 | 0.7041 | 0.7819 | 0.9178 | 0.9032 | 0.9463 | 0.8191 | 0.2702 | 0.9709 | 0.8204 | 0.8956 | 0.7080 | 0.2153 | 0.8814 |
0.2097 | 42.0 | 11424 | 0.2336 | 0.7025 | 0.7730 | 0.9180 | 0.9015 | 0.9478 | 0.8204 | 0.2303 | 0.9648 | 0.8203 | 0.8959 | 0.7089 | 0.1935 | 0.8939 |
0.1179 | 43.0 | 11696 | 0.2337 | 0.7024 | 0.7728 | 0.9181 | 0.9059 | 0.9480 | 0.8033 | 0.2422 | 0.9644 | 0.8210 | 0.8959 | 0.7061 | 0.2010 | 0.8878 |
0.1822 | 44.0 | 11968 | 0.2368 | 0.7021 | 0.7701 | 0.9181 | 0.9005 | 0.9525 | 0.7974 | 0.2376 | 0.9626 | 0.8202 | 0.8960 | 0.7055 | 0.1975 | 0.8915 |
0.1745 | 45.0 | 12240 | 0.2336 | 0.7076 | 0.7822 | 0.9182 | 0.9053 | 0.9459 | 0.8186 | 0.2732 | 0.9678 | 0.8215 | 0.8959 | 0.7097 | 0.2180 | 0.8930 |
0.1906 | 46.0 | 12512 | 0.2357 | 0.7083 | 0.7788 | 0.9185 | 0.9029 | 0.9498 | 0.8081 | 0.2745 | 0.9587 | 0.8215 | 0.8964 | 0.7080 | 0.2194 | 0.8959 |
0.188 | 47.0 | 12784 | 0.2316 | 0.7048 | 0.7765 | 0.9187 | 0.9015 | 0.9490 | 0.8201 | 0.2451 | 0.9668 | 0.8217 | 0.8966 | 0.7118 | 0.2035 | 0.8905 |
0.1621 | 48.0 | 13056 | 0.2338 | 0.7053 | 0.7789 | 0.9188 | 0.9044 | 0.9480 | 0.8156 | 0.2563 | 0.9703 | 0.8223 | 0.8966 | 0.7112 | 0.2097 | 0.8869 |
0.1239 | 49.0 | 13328 | 0.2343 | 0.7063 | 0.7786 | 0.9189 | 0.9076 | 0.9463 | 0.8151 | 0.2573 | 0.9666 | 0.8229 | 0.8966 | 0.7115 | 0.2104 | 0.8899 |
0.1934 | 50.0 | 13600 | 0.2340 | 0.7058 | 0.7765 | 0.9189 | 0.9030 | 0.9494 | 0.8141 | 0.2512 | 0.9650 | 0.8221 | 0.8966 | 0.7116 | 0.2073 | 0.8915 |
Framework versions
- Transformers 4.33.3
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3