<!-- 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-finetuned-4ss1st3r_s3gs3m-10k-steps
This model is a fine-tuned version of nvidia/mit-b0 on the blzncz/4ss1st3r_s3gs3m dataset. It achieves the following results on the evaluation set:
- Loss: 0.3966
- Mean Iou: 0.5967
- Mean Accuracy: 0.8460
- Overall Accuracy: 0.9344
- Accuracy Bg: nan
- Accuracy Fallo cohesivo: 0.9510
- Accuracy Fallo malla: 0.8524
- Accuracy Fallo adhesivo: 0.9362
- Accuracy Fallo burbuja: 0.6444
- Iou Bg: 0.0
- Iou Fallo cohesivo: 0.9239
- Iou Fallo malla: 0.7125
- Iou Fallo adhesivo: 0.8335
- Iou Fallo burbuja: 0.5139
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: 8
- eval_batch_size: 8
- seed: 1337
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: polynomial
- training_steps: 10000
Training results
Training Loss | Epoch | Step | Validation Loss | Mean Iou | Mean Accuracy | Overall Accuracy | Accuracy Bg | Accuracy Fallo cohesivo | Accuracy Fallo malla | Accuracy Fallo adhesivo | Accuracy Fallo burbuja | Iou Bg | Iou Fallo cohesivo | Iou Fallo malla | Iou Fallo adhesivo | Iou Fallo burbuja |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0.4796 | 1.0 | 133 | 0.4190 | 0.4518 | 0.6689 | 0.9049 | nan | 0.9277 | 0.8091 | 0.9381 | 0.0008 | 0.0 | 0.8866 | 0.6536 | 0.7179 | 0.0008 |
0.2665 | 2.0 | 266 | 0.3667 | 0.5096 | 0.7283 | 0.9001 | nan | 0.9111 | 0.8964 | 0.8731 | 0.2324 | 0.0 | 0.8802 | 0.6013 | 0.8467 | 0.2197 |
0.2158 | 3.0 | 399 | 0.3210 | 0.5505 | 0.7807 | 0.9142 | nan | 0.9250 | 0.8685 | 0.9414 | 0.3878 | 0.0 | 0.8952 | 0.6239 | 0.8901 | 0.3432 |
0.1737 | 4.0 | 532 | 0.3572 | 0.5370 | 0.7851 | 0.8905 | nan | 0.8905 | 0.9102 | 0.9121 | 0.4277 | 0.0 | 0.8671 | 0.5637 | 0.8777 | 0.3764 |
0.1602 | 5.0 | 665 | 0.6273 | 0.4086 | 0.7632 | 0.7743 | nan | 0.7333 | 0.9343 | 0.9685 | 0.4168 | 0.0 | 0.7198 | 0.4460 | 0.5324 | 0.3449 |
0.1707 | 6.0 | 798 | 0.3534 | 0.5442 | 0.7953 | 0.9025 | nan | 0.9056 | 0.9031 | 0.9234 | 0.4492 | 0.0 | 0.8812 | 0.5985 | 0.8629 | 0.3783 |
0.1376 | 7.0 | 931 | 0.3266 | 0.5513 | 0.7634 | 0.9262 | nan | 0.9434 | 0.8621 | 0.9288 | 0.3195 | 0.0 | 0.9109 | 0.6623 | 0.8866 | 0.2968 |
0.1346 | 8.0 | 1064 | 0.4976 | 0.4916 | 0.7900 | 0.8396 | nan | 0.8190 | 0.9133 | 0.9713 | 0.4565 | 0.0 | 0.8041 | 0.4662 | 0.7906 | 0.3970 |
0.1319 | 9.0 | 1197 | 0.3650 | 0.5652 | 0.8404 | 0.9043 | nan | 0.9053 | 0.8856 | 0.9593 | 0.6113 | 0.0 | 0.8829 | 0.5992 | 0.8734 | 0.4706 |
0.1229 | 10.0 | 1330 | 0.3201 | 0.5666 | 0.7963 | 0.9299 | nan | 0.9435 | 0.8764 | 0.9389 | 0.4265 | 0.0 | 0.9171 | 0.6896 | 0.8499 | 0.3763 |
0.1142 | 11.0 | 1463 | 0.3824 | 0.5576 | 0.8204 | 0.9020 | nan | 0.8988 | 0.9231 | 0.9456 | 0.5142 | 0.0 | 0.8795 | 0.6001 | 0.8711 | 0.4374 |
0.0983 | 12.0 | 1596 | 0.3133 | 0.5812 | 0.8297 | 0.9293 | nan | 0.9354 | 0.9046 | 0.9558 | 0.5229 | 0.0 | 0.9136 | 0.6969 | 0.8618 | 0.4335 |
0.1058 | 13.0 | 1729 | 0.2965 | 0.5860 | 0.8250 | 0.9364 | nan | 0.9528 | 0.8496 | 0.9598 | 0.5378 | 0.0 | 0.9253 | 0.7162 | 0.8502 | 0.4383 |
0.1052 | 14.0 | 1862 | 0.2839 | 0.6064 | 0.8275 | 0.9460 | nan | 0.9674 | 0.8517 | 0.9290 | 0.5621 | 0.0 | 0.9355 | 0.7492 | 0.8930 | 0.4540 |
0.0911 | 15.0 | 1995 | 0.3245 | 0.5853 | 0.8116 | 0.9368 | nan | 0.9565 | 0.8504 | 0.9298 | 0.5099 | 0.0 | 0.9243 | 0.7171 | 0.8534 | 0.4318 |
0.0889 | 16.0 | 2128 | 0.3094 | 0.5969 | 0.8225 | 0.9422 | nan | 0.9615 | 0.8559 | 0.9376 | 0.5351 | 0.0 | 0.9313 | 0.7353 | 0.8726 | 0.4451 |
0.0827 | 17.0 | 2261 | 0.4776 | 0.5187 | 0.8195 | 0.8547 | nan | 0.8390 | 0.9163 | 0.9440 | 0.5786 | 0.0 | 0.8207 | 0.4920 | 0.8216 | 0.4590 |
0.0939 | 18.0 | 2394 | 0.3923 | 0.5364 | 0.8375 | 0.8948 | nan | 0.8950 | 0.8831 | 0.9437 | 0.6282 | 0.0 | 0.8746 | 0.6268 | 0.7090 | 0.4717 |
0.0799 | 19.0 | 2527 | 0.3560 | 0.5776 | 0.8252 | 0.9254 | nan | 0.9337 | 0.8933 | 0.9409 | 0.5331 | 0.0 | 0.9096 | 0.6846 | 0.8519 | 0.4422 |
0.075 | 20.0 | 2660 | 0.3803 | 0.5796 | 0.8338 | 0.9194 | nan | 0.9249 | 0.9078 | 0.9238 | 0.5788 | 0.0 | 0.9032 | 0.6459 | 0.8821 | 0.4670 |
0.0844 | 21.0 | 2793 | 0.2885 | 0.6170 | 0.8334 | 0.9507 | nan | 0.9757 | 0.8296 | 0.9390 | 0.5892 | 0.0 | 0.9412 | 0.7654 | 0.8933 | 0.4852 |
0.0746 | 22.0 | 2926 | 0.3222 | 0.5831 | 0.8160 | 0.9331 | nan | 0.9481 | 0.8685 | 0.9370 | 0.5105 | 0.0 | 0.9193 | 0.7032 | 0.8716 | 0.4215 |
0.072 | 23.0 | 3059 | 0.3481 | 0.5878 | 0.8336 | 0.9266 | nan | 0.9357 | 0.8952 | 0.9271 | 0.5764 | 0.0 | 0.9123 | 0.6824 | 0.8720 | 0.4725 |
0.0735 | 24.0 | 3192 | 0.3196 | 0.5974 | 0.8403 | 0.9353 | nan | 0.9496 | 0.8666 | 0.9430 | 0.6018 | 0.0 | 0.9225 | 0.7165 | 0.8649 | 0.4832 |
0.0674 | 25.0 | 3325 | 0.3407 | 0.5927 | 0.8435 | 0.9282 | nan | 0.9401 | 0.8786 | 0.9246 | 0.6304 | 0.0 | 0.9141 | 0.6844 | 0.8696 | 0.4953 |
0.0712 | 26.0 | 3458 | 0.3356 | 0.5906 | 0.8420 | 0.9301 | nan | 0.9405 | 0.8895 | 0.9299 | 0.6080 | 0.0 | 0.9160 | 0.6905 | 0.8743 | 0.4722 |
0.072 | 27.0 | 3591 | 0.3491 | 0.5833 | 0.8372 | 0.9286 | nan | 0.9415 | 0.8636 | 0.9425 | 0.6012 | 0.0 | 0.9161 | 0.6966 | 0.8246 | 0.4790 |
0.0641 | 28.0 | 3724 | 0.3130 | 0.6087 | 0.8422 | 0.9473 | nan | 0.9697 | 0.8357 | 0.9427 | 0.6208 | 0.0 | 0.9386 | 0.7613 | 0.8599 | 0.4837 |
0.0597 | 29.0 | 3857 | 0.3828 | 0.5666 | 0.8394 | 0.9107 | nan | 0.9141 | 0.8934 | 0.9411 | 0.6092 | 0.0 | 0.8924 | 0.6327 | 0.8343 | 0.4735 |
0.0648 | 30.0 | 3990 | 0.3435 | 0.6001 | 0.8372 | 0.9403 | nan | 0.9569 | 0.8708 | 0.9276 | 0.5935 | 0.0 | 0.9292 | 0.7312 | 0.8779 | 0.4623 |
0.0618 | 31.0 | 4123 | 0.3531 | 0.5963 | 0.8521 | 0.9303 | nan | 0.9450 | 0.8621 | 0.9240 | 0.6773 | 0.0 | 0.9179 | 0.6842 | 0.8730 | 0.5063 |
0.0556 | 32.0 | 4256 | 0.3307 | 0.6037 | 0.8417 | 0.9401 | nan | 0.9576 | 0.8637 | 0.9271 | 0.6183 | 0.0 | 0.9298 | 0.7274 | 0.8637 | 0.4974 |
0.0616 | 33.0 | 4389 | 0.3510 | 0.5911 | 0.8347 | 0.9298 | nan | 0.9424 | 0.8714 | 0.9388 | 0.5863 | 0.0 | 0.9158 | 0.6914 | 0.8745 | 0.4740 |
0.0603 | 34.0 | 4522 | 0.3467 | 0.6022 | 0.8544 | 0.9334 | nan | 0.9487 | 0.8610 | 0.9274 | 0.6807 | 0.0 | 0.9211 | 0.7029 | 0.8738 | 0.5130 |
0.0587 | 35.0 | 4655 | 0.3574 | 0.6017 | 0.8407 | 0.9379 | nan | 0.9555 | 0.8541 | 0.9346 | 0.6187 | 0.0 | 0.9269 | 0.7228 | 0.8627 | 0.4962 |
0.0557 | 36.0 | 4788 | 0.3871 | 0.5720 | 0.8334 | 0.9178 | nan | 0.9317 | 0.8416 | 0.9374 | 0.6228 | 0.0 | 0.9051 | 0.6479 | 0.8160 | 0.4911 |
0.0567 | 37.0 | 4921 | 0.4425 | 0.5656 | 0.8282 | 0.9070 | nan | 0.9114 | 0.8922 | 0.9244 | 0.5848 | 0.0 | 0.8889 | 0.6100 | 0.8575 | 0.4718 |
0.0537 | 38.0 | 5054 | 0.3512 | 0.5946 | 0.8392 | 0.9317 | nan | 0.9463 | 0.8649 | 0.9314 | 0.6142 | 0.0 | 0.9187 | 0.6984 | 0.8637 | 0.4921 |
0.0559 | 39.0 | 5187 | 0.3676 | 0.5931 | 0.8437 | 0.9273 | nan | 0.9381 | 0.8798 | 0.9323 | 0.6247 | 0.0 | 0.9129 | 0.6779 | 0.8786 | 0.4959 |
0.0502 | 40.0 | 5320 | 0.4149 | 0.5518 | 0.8381 | 0.8984 | nan | 0.9011 | 0.8773 | 0.9368 | 0.6370 | 0.0 | 0.8793 | 0.6069 | 0.7741 | 0.4989 |
0.0559 | 41.0 | 5453 | 0.4042 | 0.5694 | 0.8342 | 0.9130 | nan | 0.9206 | 0.8721 | 0.9400 | 0.6041 | 0.0 | 0.8971 | 0.6319 | 0.8286 | 0.4896 |
0.0523 | 42.0 | 5586 | 0.3669 | 0.5903 | 0.8462 | 0.9286 | nan | 0.9414 | 0.8676 | 0.9337 | 0.6421 | 0.0 | 0.9162 | 0.6883 | 0.8370 | 0.5102 |
0.0525 | 43.0 | 5719 | 0.4140 | 0.5704 | 0.8531 | 0.9081 | nan | 0.9110 | 0.8867 | 0.9417 | 0.6729 | 0.0 | 0.8898 | 0.6220 | 0.8366 | 0.5035 |
0.0508 | 44.0 | 5852 | 0.3965 | 0.5714 | 0.8396 | 0.9141 | nan | 0.9227 | 0.8800 | 0.9147 | 0.6409 | 0.0 | 0.8989 | 0.6513 | 0.8007 | 0.5060 |
0.0507 | 45.0 | 5985 | 0.3793 | 0.5817 | 0.8392 | 0.9196 | nan | 0.9272 | 0.8932 | 0.9214 | 0.6148 | 0.0 | 0.9042 | 0.6627 | 0.8407 | 0.5011 |
0.0494 | 46.0 | 6118 | 0.3500 | 0.6020 | 0.8426 | 0.9363 | nan | 0.9524 | 0.8619 | 0.9322 | 0.6240 | 0.0 | 0.9247 | 0.7142 | 0.8653 | 0.5058 |
0.0462 | 47.0 | 6251 | 0.3524 | 0.6031 | 0.8435 | 0.9388 | nan | 0.9545 | 0.8668 | 0.9364 | 0.6163 | 0.0 | 0.9274 | 0.7269 | 0.8703 | 0.4909 |
0.0486 | 48.0 | 6384 | 0.3876 | 0.5902 | 0.8397 | 0.9308 | nan | 0.9479 | 0.8557 | 0.9161 | 0.6392 | 0.0 | 0.9203 | 0.6928 | 0.8334 | 0.5046 |
0.0461 | 49.0 | 6517 | 0.3674 | 0.5933 | 0.8409 | 0.9326 | nan | 0.9482 | 0.8622 | 0.9258 | 0.6274 | 0.0 | 0.9214 | 0.7053 | 0.8367 | 0.5030 |
0.0497 | 50.0 | 6650 | 0.4018 | 0.5838 | 0.8374 | 0.9246 | nan | 0.9390 | 0.8519 | 0.9341 | 0.6244 | 0.0 | 0.9102 | 0.6733 | 0.8361 | 0.4992 |
0.0491 | 51.0 | 6783 | 0.4036 | 0.5824 | 0.8513 | 0.9198 | nan | 0.9272 | 0.8805 | 0.9403 | 0.6573 | 0.0 | 0.9037 | 0.6712 | 0.8169 | 0.5203 |
0.046 | 52.0 | 6916 | 0.3913 | 0.5820 | 0.8395 | 0.9243 | nan | 0.9347 | 0.8771 | 0.9336 | 0.6126 | 0.0 | 0.9105 | 0.6792 | 0.8244 | 0.4960 |
0.0488 | 53.0 | 7049 | 0.3441 | 0.6010 | 0.8504 | 0.9362 | nan | 0.9523 | 0.8521 | 0.9457 | 0.6517 | 0.0 | 0.9250 | 0.7225 | 0.8496 | 0.5081 |
0.0458 | 54.0 | 7182 | 0.3784 | 0.5977 | 0.8382 | 0.9378 | nan | 0.9603 | 0.8212 | 0.9375 | 0.6337 | 0.0 | 0.9286 | 0.7157 | 0.8387 | 0.5053 |
0.0449 | 55.0 | 7315 | 0.3506 | 0.6068 | 0.8493 | 0.9404 | nan | 0.9579 | 0.8554 | 0.9385 | 0.6456 | 0.0 | 0.9300 | 0.7357 | 0.8549 | 0.5132 |
0.0482 | 56.0 | 7448 | 0.4005 | 0.5819 | 0.8414 | 0.9249 | nan | 0.9374 | 0.8642 | 0.9337 | 0.6303 | 0.0 | 0.9119 | 0.6831 | 0.8139 | 0.5006 |
0.0434 | 57.0 | 7581 | 0.3749 | 0.5914 | 0.8465 | 0.9294 | nan | 0.9423 | 0.8675 | 0.9339 | 0.6421 | 0.0 | 0.9171 | 0.6999 | 0.8265 | 0.5134 |
0.0435 | 58.0 | 7714 | 0.4195 | 0.5722 | 0.8400 | 0.9172 | nan | 0.9274 | 0.8700 | 0.9234 | 0.6392 | 0.0 | 0.9025 | 0.6588 | 0.7954 | 0.5044 |
0.0442 | 59.0 | 7847 | 0.3975 | 0.5828 | 0.8407 | 0.9257 | nan | 0.9398 | 0.8563 | 0.9312 | 0.6356 | 0.0 | 0.9134 | 0.6866 | 0.8103 | 0.5037 |
0.0442 | 60.0 | 7980 | 0.3845 | 0.5929 | 0.8457 | 0.9315 | nan | 0.9459 | 0.8603 | 0.9363 | 0.6404 | 0.0 | 0.9193 | 0.7041 | 0.8308 | 0.5103 |
0.0422 | 61.0 | 8113 | 0.3875 | 0.5963 | 0.8465 | 0.9338 | nan | 0.9489 | 0.8616 | 0.9340 | 0.6413 | 0.0 | 0.9226 | 0.7135 | 0.8381 | 0.5072 |
0.0436 | 62.0 | 8246 | 0.3859 | 0.6022 | 0.8497 | 0.9385 | nan | 0.9566 | 0.8477 | 0.9382 | 0.6562 | 0.0 | 0.9289 | 0.7300 | 0.8376 | 0.5147 |
0.0429 | 63.0 | 8379 | 0.3857 | 0.5956 | 0.8425 | 0.9357 | nan | 0.9534 | 0.8481 | 0.9357 | 0.6327 | 0.0 | 0.9249 | 0.7233 | 0.8283 | 0.5016 |
0.0446 | 64.0 | 8512 | 0.3778 | 0.5976 | 0.8495 | 0.9343 | nan | 0.9492 | 0.8602 | 0.9399 | 0.6489 | 0.0 | 0.9232 | 0.7191 | 0.8305 | 0.5153 |
0.0429 | 65.0 | 8645 | 0.3889 | 0.5948 | 0.8478 | 0.9330 | nan | 0.9490 | 0.8548 | 0.9325 | 0.6549 | 0.0 | 0.9225 | 0.7075 | 0.8271 | 0.5167 |
0.0454 | 66.0 | 8778 | 0.3915 | 0.5941 | 0.8470 | 0.9329 | nan | 0.9490 | 0.8571 | 0.9271 | 0.6547 | 0.0 | 0.9221 | 0.7087 | 0.8278 | 0.5117 |
0.0427 | 67.0 | 8911 | 0.3924 | 0.5967 | 0.8455 | 0.9349 | nan | 0.9518 | 0.8520 | 0.9350 | 0.6433 | 0.0 | 0.9247 | 0.7167 | 0.8290 | 0.5133 |
0.0425 | 68.0 | 9044 | 0.3990 | 0.5992 | 0.8491 | 0.9358 | nan | 0.9524 | 0.8545 | 0.9355 | 0.6541 | 0.0 | 0.9250 | 0.7187 | 0.8387 | 0.5136 |
0.0429 | 69.0 | 9177 | 0.3911 | 0.5909 | 0.8499 | 0.9303 | nan | 0.9451 | 0.8532 | 0.9394 | 0.6619 | 0.0 | 0.9192 | 0.7029 | 0.8178 | 0.5146 |
0.0465 | 70.0 | 9310 | 0.3840 | 0.5977 | 0.8481 | 0.9332 | nan | 0.9473 | 0.8700 | 0.9278 | 0.6473 | 0.0 | 0.9215 | 0.7079 | 0.8480 | 0.5110 |
0.0436 | 71.0 | 9443 | 0.3862 | 0.5974 | 0.8456 | 0.9351 | nan | 0.9518 | 0.8534 | 0.9359 | 0.6413 | 0.0 | 0.9248 | 0.7162 | 0.8338 | 0.5124 |
0.0435 | 72.0 | 9576 | 0.3926 | 0.5952 | 0.8448 | 0.9328 | nan | 0.9484 | 0.8585 | 0.9318 | 0.6405 | 0.0 | 0.9217 | 0.7073 | 0.8386 | 0.5084 |
0.0421 | 73.0 | 9709 | 0.3961 | 0.5984 | 0.8467 | 0.9348 | nan | 0.9513 | 0.8564 | 0.9309 | 0.6482 | 0.0 | 0.9243 | 0.7119 | 0.8414 | 0.5143 |
0.0409 | 74.0 | 9842 | 0.3973 | 0.5982 | 0.8494 | 0.9341 | nan | 0.9498 | 0.8596 | 0.9306 | 0.6578 | 0.0 | 0.9233 | 0.7094 | 0.8401 | 0.5181 |
0.041 | 75.0 | 9975 | 0.3898 | 0.5963 | 0.8476 | 0.9335 | nan | 0.9493 | 0.8561 | 0.9354 | 0.6498 | 0.0 | 0.9227 | 0.7108 | 0.8329 | 0.5153 |
0.0436 | 75.19 | 10000 | 0.3966 | 0.5967 | 0.8460 | 0.9344 | nan | 0.9510 | 0.8524 | 0.9362 | 0.6444 | 0.0 | 0.9239 | 0.7125 | 0.8335 | 0.5139 |
Framework versions
- Transformers 4.31.0.dev0
- Pytorch 2.0.1+cpu
- Datasets 2.13.1
- Tokenizers 0.13.3