segformer-b0_DsMetalDam_Train_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: 0.2158
- Mean Iou: 0.7158
- Mean Accuracy: 0.7885
- Overall Accuracy: 0.9224
- Accuracy Matrix: 0.9019
- Accuracy Austenite: 0.9539
- Accuracy Martensite/austenite: 0.8174
- Accuracy Precipitate: 0.3067
- Accuracy Defect: 0.9625
- Iou Matrix: 0.8320
- Iou Austenite: 0.8988
- Iou Martensite/austenite: 0.7096
- Iou Precipitate: 0.2363
- Iou Defect: 0.9024
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 |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0.4826 | 1.0 | 327 | 0.4735 | 0.5267 | 0.6425 | 0.8330 | 0.7351 | 0.9126 | 0.7551 | 0.0 | 0.8096 | 0.6604 | 0.8334 | 0.4020 | 0.0 | 0.7378 |
0.5204 | 2.0 | 654 | 0.3567 | 0.5622 | 0.6354 | 0.8720 | 0.8799 | 0.9316 | 0.4736 | 0.0 | 0.8919 | 0.7422 | 0.8521 | 0.4239 | 0.0 | 0.7931 |
0.1559 | 3.0 | 981 | 0.3234 | 0.5879 | 0.6646 | 0.8797 | 0.8867 | 0.9260 | 0.5670 | 0.0 | 0.9436 | 0.7525 | 0.8583 | 0.4930 | 0.0 | 0.8354 |
0.2957 | 4.0 | 1308 | 0.3037 | 0.5972 | 0.6781 | 0.8862 | 0.8723 | 0.9393 | 0.6163 | 0.0 | 0.9626 | 0.7625 | 0.8649 | 0.5289 | 0.0 | 0.8298 |
0.2695 | 5.0 | 1635 | 0.2920 | 0.6007 | 0.6923 | 0.8890 | 0.8703 | 0.9338 | 0.6929 | 0.0 | 0.9646 | 0.7681 | 0.8680 | 0.5577 | 0.0 | 0.8097 |
0.7461 | 6.0 | 1962 | 0.2805 | 0.6137 | 0.7045 | 0.8923 | 0.8600 | 0.9370 | 0.7549 | 0.0 | 0.9703 | 0.7724 | 0.8711 | 0.5850 | 0.0 | 0.8397 |
0.1637 | 7.0 | 2289 | 0.2660 | 0.6245 | 0.6967 | 0.8982 | 0.8940 | 0.9342 | 0.7024 | 0.0013 | 0.9516 | 0.7861 | 0.8753 | 0.5952 | 0.0013 | 0.8648 |
0.1495 | 8.0 | 2616 | 0.2676 | 0.6240 | 0.7089 | 0.8966 | 0.8657 | 0.9373 | 0.7813 | 0.0052 | 0.9550 | 0.7796 | 0.8760 | 0.6039 | 0.0052 | 0.8555 |
0.2658 | 9.0 | 2943 | 0.2620 | 0.6345 | 0.7127 | 0.9005 | 0.8747 | 0.9403 | 0.7699 | 0.0257 | 0.9529 | 0.7886 | 0.8772 | 0.6204 | 0.0250 | 0.8613 |
0.2517 | 10.0 | 3270 | 0.2532 | 0.6529 | 0.7218 | 0.9040 | 0.8884 | 0.9394 | 0.7607 | 0.0885 | 0.9321 | 0.7964 | 0.8797 | 0.6333 | 0.0799 | 0.8753 |
0.152 | 11.0 | 3597 | 0.2490 | 0.6606 | 0.7273 | 0.9052 | 0.8820 | 0.9475 | 0.7492 | 0.1412 | 0.9165 | 0.7971 | 0.8812 | 0.6373 | 0.1222 | 0.8652 |
0.4032 | 12.0 | 3924 | 0.2470 | 0.6606 | 0.7329 | 0.9062 | 0.8873 | 0.9406 | 0.7806 | 0.1045 | 0.9516 | 0.7997 | 0.8821 | 0.6464 | 0.0954 | 0.8796 |
0.1497 | 13.0 | 4251 | 0.2479 | 0.6608 | 0.7334 | 0.9069 | 0.8896 | 0.9415 | 0.7729 | 0.1016 | 0.9615 | 0.8015 | 0.8823 | 0.6487 | 0.0917 | 0.8797 |
0.2648 | 14.0 | 4578 | 0.2441 | 0.6580 | 0.7258 | 0.9081 | 0.8954 | 0.9399 | 0.7758 | 0.0742 | 0.9434 | 0.8048 | 0.8836 | 0.6521 | 0.0654 | 0.8841 |
0.2517 | 15.0 | 4905 | 0.2389 | 0.6858 | 0.7683 | 0.9101 | 0.8936 | 0.9468 | 0.7543 | 0.2911 | 0.9555 | 0.8077 | 0.8863 | 0.6557 | 0.1957 | 0.8838 |
0.4088 | 16.0 | 5232 | 0.2344 | 0.6849 | 0.7577 | 0.9116 | 0.8983 | 0.9425 | 0.7828 | 0.2164 | 0.9488 | 0.8104 | 0.8872 | 0.6676 | 0.1772 | 0.8821 |
0.7573 | 17.0 | 5559 | 0.2329 | 0.6893 | 0.7678 | 0.9127 | 0.8986 | 0.9443 | 0.7803 | 0.2521 | 0.9634 | 0.8126 | 0.8887 | 0.6713 | 0.1922 | 0.8818 |
0.361 | 18.0 | 5886 | 0.2352 | 0.6844 | 0.7547 | 0.9129 | 0.9003 | 0.9458 | 0.7686 | 0.2055 | 0.9534 | 0.8136 | 0.8890 | 0.6658 | 0.1672 | 0.8865 |
0.0933 | 19.0 | 6213 | 0.2302 | 0.6929 | 0.7691 | 0.9132 | 0.8960 | 0.9423 | 0.8121 | 0.2513 | 0.9435 | 0.8143 | 0.8891 | 0.6750 | 0.1950 | 0.8910 |
0.2342 | 20.0 | 6540 | 0.2325 | 0.6805 | 0.7436 | 0.9134 | 0.8885 | 0.9530 | 0.7799 | 0.1512 | 0.9456 | 0.8130 | 0.8886 | 0.6758 | 0.1313 | 0.8936 |
0.0977 | 21.0 | 6867 | 0.2303 | 0.6927 | 0.7679 | 0.9147 | 0.9033 | 0.9437 | 0.7897 | 0.2524 | 0.9502 | 0.8178 | 0.8902 | 0.6791 | 0.1890 | 0.8876 |
0.144 | 22.0 | 7194 | 0.2277 | 0.6921 | 0.7674 | 0.9146 | 0.8850 | 0.9539 | 0.7986 | 0.2441 | 0.9555 | 0.8149 | 0.8913 | 0.6788 | 0.1885 | 0.8872 |
0.093 | 23.0 | 7521 | 0.2298 | 0.6911 | 0.7612 | 0.9154 | 0.9035 | 0.9431 | 0.8014 | 0.2015 | 0.9564 | 0.8192 | 0.8910 | 0.6811 | 0.1643 | 0.8998 |
0.1885 | 24.0 | 7848 | 0.2231 | 0.6910 | 0.7651 | 0.9166 | 0.8901 | 0.9513 | 0.8184 | 0.1993 | 0.9667 | 0.8190 | 0.8933 | 0.6910 | 0.1647 | 0.8868 |
0.1533 | 25.0 | 8175 | 0.2208 | 0.6946 | 0.7786 | 0.9169 | 0.8939 | 0.9477 | 0.8263 | 0.2488 | 0.9765 | 0.8202 | 0.8937 | 0.6930 | 0.1962 | 0.8698 |
0.1924 | 26.0 | 8502 | 0.2250 | 0.6989 | 0.7688 | 0.9173 | 0.9060 | 0.9462 | 0.7923 | 0.2436 | 0.9558 | 0.8219 | 0.8941 | 0.6853 | 0.1948 | 0.8983 |
0.1499 | 27.0 | 8829 | 0.2227 | 0.7001 | 0.7698 | 0.9178 | 0.8955 | 0.9508 | 0.8132 | 0.2381 | 0.9511 | 0.8224 | 0.8939 | 0.6937 | 0.1901 | 0.9003 |
0.2098 | 28.0 | 9156 | 0.2212 | 0.7053 | 0.7788 | 0.9181 | 0.8947 | 0.9519 | 0.8127 | 0.2950 | 0.9395 | 0.8227 | 0.8947 | 0.6942 | 0.2184 | 0.8967 |
0.0973 | 29.0 | 9483 | 0.2181 | 0.6930 | 0.7643 | 0.9192 | 0.9014 | 0.9476 | 0.8256 | 0.1744 | 0.9723 | 0.8248 | 0.8957 | 0.7018 | 0.1574 | 0.8851 |
0.1031 | 30.0 | 9810 | 0.2197 | 0.7056 | 0.7822 | 0.9188 | 0.9089 | 0.9440 | 0.8100 | 0.2872 | 0.9610 | 0.8257 | 0.8946 | 0.6971 | 0.2147 | 0.8961 |
0.0886 | 31.0 | 10137 | 0.2221 | 0.7089 | 0.7898 | 0.9187 | 0.8989 | 0.9498 | 0.8115 | 0.3236 | 0.9653 | 0.8248 | 0.8948 | 0.6962 | 0.2319 | 0.8966 |
0.1318 | 32.0 | 10464 | 0.2216 | 0.7048 | 0.7767 | 0.9194 | 0.9103 | 0.9463 | 0.7981 | 0.2698 | 0.9590 | 0.8272 | 0.8949 | 0.6981 | 0.2085 | 0.8954 |
0.1206 | 33.0 | 10791 | 0.2188 | 0.7082 | 0.7860 | 0.9196 | 0.8973 | 0.9502 | 0.8280 | 0.2927 | 0.9617 | 0.8264 | 0.8964 | 0.6989 | 0.2195 | 0.8999 |
0.2152 | 34.0 | 11118 | 0.2179 | 0.7112 | 0.7846 | 0.9198 | 0.8937 | 0.9537 | 0.8234 | 0.2997 | 0.9523 | 0.8254 | 0.8963 | 0.7034 | 0.2279 | 0.9029 |
0.1581 | 35.0 | 11445 | 0.2170 | 0.7083 | 0.7839 | 0.9204 | 0.9043 | 0.9495 | 0.8116 | 0.2810 | 0.9729 | 0.8278 | 0.8970 | 0.7021 | 0.2257 | 0.8890 |
0.3181 | 36.0 | 11772 | 0.2185 | 0.7097 | 0.7850 | 0.9202 | 0.9128 | 0.9443 | 0.8080 | 0.2957 | 0.9641 | 0.8289 | 0.8960 | 0.7016 | 0.2241 | 0.8977 |
0.3164 | 37.0 | 12099 | 0.2172 | 0.7135 | 0.7907 | 0.9208 | 0.9103 | 0.9470 | 0.8074 | 0.3300 | 0.9587 | 0.8300 | 0.8969 | 0.7023 | 0.2379 | 0.9003 |
0.1082 | 38.0 | 12426 | 0.2180 | 0.7131 | 0.7859 | 0.9207 | 0.8981 | 0.9530 | 0.8199 | 0.3055 | 0.9530 | 0.8283 | 0.8973 | 0.7029 | 0.2345 | 0.9023 |
0.1884 | 39.0 | 12753 | 0.2187 | 0.7115 | 0.7851 | 0.9207 | 0.9010 | 0.9508 | 0.8223 | 0.2911 | 0.9604 | 0.8291 | 0.8969 | 0.7038 | 0.2256 | 0.9018 |
0.1169 | 40.0 | 13080 | 0.2171 | 0.7130 | 0.7858 | 0.9215 | 0.9066 | 0.9497 | 0.8146 | 0.2965 | 0.9616 | 0.8309 | 0.8978 | 0.7051 | 0.2286 | 0.9026 |
0.1622 | 41.0 | 13407 | 0.2160 | 0.7114 | 0.7808 | 0.9215 | 0.9020 | 0.9512 | 0.8260 | 0.2715 | 0.9532 | 0.8304 | 0.8979 | 0.7068 | 0.2183 | 0.9035 |
0.2682 | 42.0 | 13734 | 0.2164 | 0.7120 | 0.7840 | 0.9219 | 0.9054 | 0.9496 | 0.8255 | 0.2751 | 0.9647 | 0.8313 | 0.8985 | 0.7077 | 0.2206 | 0.9018 |
0.1214 | 43.0 | 14061 | 0.2159 | 0.7044 | 0.7717 | 0.9217 | 0.9083 | 0.9485 | 0.8199 | 0.2176 | 0.9641 | 0.8315 | 0.8978 | 0.7072 | 0.1835 | 0.9020 |
0.1344 | 44.0 | 14388 | 0.2153 | 0.7145 | 0.7906 | 0.9220 | 0.9039 | 0.9506 | 0.8251 | 0.3057 | 0.9677 | 0.8318 | 0.8985 | 0.7080 | 0.2338 | 0.9006 |
0.1233 | 45.0 | 14715 | 0.2142 | 0.7135 | 0.7859 | 0.9220 | 0.9063 | 0.9488 | 0.8285 | 0.2853 | 0.9606 | 0.8319 | 0.8986 | 0.7084 | 0.2253 | 0.9030 |
0.1819 | 46.0 | 15042 | 0.2150 | 0.7156 | 0.7878 | 0.9224 | 0.9058 | 0.9504 | 0.8244 | 0.2985 | 0.9601 | 0.8324 | 0.8989 | 0.7098 | 0.2327 | 0.9042 |
0.1338 | 47.0 | 15369 | 0.2159 | 0.7130 | 0.7846 | 0.9223 | 0.9080 | 0.9502 | 0.8151 | 0.2853 | 0.9646 | 0.8328 | 0.8987 | 0.7072 | 0.2247 | 0.9015 |
0.1299 | 48.0 | 15696 | 0.2129 | 0.7158 | 0.7882 | 0.9224 | 0.9017 | 0.9524 | 0.8288 | 0.2965 | 0.9617 | 0.8319 | 0.8989 | 0.7113 | 0.2336 | 0.9030 |
0.1295 | 49.0 | 16023 | 0.2163 | 0.7147 | 0.7854 | 0.9224 | 0.9120 | 0.9477 | 0.8157 | 0.2948 | 0.9567 | 0.8332 | 0.8986 | 0.7081 | 0.2296 | 0.9042 |
0.1567 | 50.0 | 16350 | 0.2158 | 0.7158 | 0.7885 | 0.9224 | 0.9019 | 0.9539 | 0.8174 | 0.3067 | 0.9625 | 0.8320 | 0.8988 | 0.7096 | 0.2363 | 0.9024 |
Framework versions
- Transformers 4.33.2
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
- Downloads last month
- 0
Model tree for ironchanchellor/segformer-b0_DsMetalDam_Train_Augmented_Cropped
Base model
nvidia/mit-b0