<!-- 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. -->
vit-tiny_rvl_cdip_100_examples_per_class_kd_CEKD_t1.5_a0.9
This model is a fine-tuned version of WinKawaks/vit-tiny-patch16-224 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.5908
- Accuracy: 0.54
- Brier Loss: 0.6121
- Nll: 2.4999
- F1 Micro: 0.54
- F1 Macro: 0.5334
- Ece: 0.1895
- Aurc: 0.2228
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: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 100
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
---|---|---|---|---|---|---|---|---|---|---|
No log | 1.0 | 7 | 3.9334 | 0.0425 | 1.0719 | 7.3494 | 0.0425 | 0.0341 | 0.2781 | 0.9557 |
No log | 2.0 | 14 | 3.0206 | 0.09 | 0.9526 | 5.9271 | 0.09 | 0.0682 | 0.1659 | 0.8896 |
No log | 3.0 | 21 | 2.6955 | 0.2275 | 0.8867 | 5.4041 | 0.2275 | 0.1804 | 0.1637 | 0.6351 |
No log | 4.0 | 28 | 2.3029 | 0.29 | 0.8005 | 3.5699 | 0.29 | 0.2727 | 0.1512 | 0.4877 |
No log | 5.0 | 35 | 2.0293 | 0.37 | 0.7219 | 2.9856 | 0.37 | 0.3519 | 0.1495 | 0.3632 |
No log | 6.0 | 42 | 1.9194 | 0.44 | 0.6926 | 2.9746 | 0.44 | 0.4240 | 0.1778 | 0.3091 |
No log | 7.0 | 49 | 1.9640 | 0.4675 | 0.7000 | 3.0857 | 0.4675 | 0.4390 | 0.2009 | 0.3004 |
No log | 8.0 | 56 | 1.9049 | 0.4625 | 0.7003 | 2.8472 | 0.4625 | 0.4602 | 0.2053 | 0.3051 |
No log | 9.0 | 63 | 2.0561 | 0.4675 | 0.7168 | 2.8229 | 0.4675 | 0.4501 | 0.2288 | 0.2994 |
No log | 10.0 | 70 | 2.1002 | 0.45 | 0.7433 | 2.7915 | 0.45 | 0.4234 | 0.2691 | 0.3022 |
No log | 11.0 | 77 | 2.2528 | 0.4525 | 0.7686 | 3.0103 | 0.4525 | 0.4320 | 0.2921 | 0.3183 |
No log | 12.0 | 84 | 2.1190 | 0.475 | 0.7427 | 2.6715 | 0.4750 | 0.4660 | 0.2832 | 0.3077 |
No log | 13.0 | 91 | 2.3102 | 0.445 | 0.7825 | 2.9698 | 0.445 | 0.4252 | 0.3093 | 0.3100 |
No log | 14.0 | 98 | 2.3501 | 0.42 | 0.8145 | 2.7585 | 0.4200 | 0.4248 | 0.3206 | 0.3662 |
No log | 15.0 | 105 | 2.2402 | 0.495 | 0.7423 | 3.0313 | 0.495 | 0.4702 | 0.2692 | 0.2818 |
No log | 16.0 | 112 | 2.2266 | 0.49 | 0.7349 | 2.8824 | 0.49 | 0.4714 | 0.2763 | 0.2895 |
No log | 17.0 | 119 | 2.2989 | 0.4725 | 0.7509 | 3.0951 | 0.4725 | 0.4499 | 0.2863 | 0.2855 |
No log | 18.0 | 126 | 2.1355 | 0.47 | 0.7322 | 2.9349 | 0.47 | 0.4616 | 0.2725 | 0.2845 |
No log | 19.0 | 133 | 2.0965 | 0.505 | 0.7067 | 2.8254 | 0.505 | 0.4956 | 0.2523 | 0.2757 |
No log | 20.0 | 140 | 2.1961 | 0.485 | 0.7358 | 3.1604 | 0.485 | 0.4567 | 0.2825 | 0.2841 |
No log | 21.0 | 147 | 2.1287 | 0.5025 | 0.7247 | 2.5998 | 0.5025 | 0.5074 | 0.2703 | 0.3064 |
No log | 22.0 | 154 | 2.2280 | 0.4675 | 0.7760 | 2.8571 | 0.4675 | 0.4636 | 0.2911 | 0.3232 |
No log | 23.0 | 161 | 1.9649 | 0.5025 | 0.6828 | 2.8224 | 0.5025 | 0.4970 | 0.2410 | 0.2633 |
No log | 24.0 | 168 | 1.9361 | 0.5125 | 0.6780 | 2.7309 | 0.5125 | 0.5035 | 0.2326 | 0.2553 |
No log | 25.0 | 175 | 2.0161 | 0.5 | 0.6980 | 2.9958 | 0.5 | 0.4912 | 0.2580 | 0.2556 |
No log | 26.0 | 182 | 1.8763 | 0.5025 | 0.6624 | 2.8291 | 0.5025 | 0.4952 | 0.2305 | 0.2431 |
No log | 27.0 | 189 | 1.9057 | 0.525 | 0.6793 | 2.5627 | 0.525 | 0.5174 | 0.2161 | 0.2634 |
No log | 28.0 | 196 | 1.8529 | 0.52 | 0.6683 | 2.7191 | 0.52 | 0.5132 | 0.2375 | 0.2535 |
No log | 29.0 | 203 | 1.9603 | 0.5125 | 0.6831 | 2.7822 | 0.5125 | 0.5076 | 0.2395 | 0.2657 |
No log | 30.0 | 210 | 1.8247 | 0.52 | 0.6533 | 2.8547 | 0.52 | 0.5058 | 0.2080 | 0.2426 |
No log | 31.0 | 217 | 1.8275 | 0.5125 | 0.6547 | 2.6194 | 0.5125 | 0.5032 | 0.2208 | 0.2488 |
No log | 32.0 | 224 | 1.8003 | 0.52 | 0.6455 | 2.6138 | 0.52 | 0.5124 | 0.2302 | 0.2370 |
No log | 33.0 | 231 | 1.8714 | 0.505 | 0.6694 | 2.6643 | 0.505 | 0.4970 | 0.2195 | 0.2553 |
No log | 34.0 | 238 | 1.8018 | 0.5075 | 0.6659 | 2.5423 | 0.5075 | 0.4978 | 0.2241 | 0.2515 |
No log | 35.0 | 245 | 1.7844 | 0.5225 | 0.6503 | 2.6100 | 0.5225 | 0.5181 | 0.2181 | 0.2435 |
No log | 36.0 | 252 | 1.8321 | 0.5225 | 0.6674 | 2.7821 | 0.5225 | 0.5020 | 0.2285 | 0.2462 |
No log | 37.0 | 259 | 1.7859 | 0.4975 | 0.6725 | 2.6066 | 0.4975 | 0.4974 | 0.2351 | 0.2627 |
No log | 38.0 | 266 | 1.7790 | 0.5125 | 0.6595 | 2.6983 | 0.5125 | 0.5023 | 0.2172 | 0.2497 |
No log | 39.0 | 273 | 1.6989 | 0.5225 | 0.6401 | 2.6743 | 0.5225 | 0.5151 | 0.2100 | 0.2407 |
No log | 40.0 | 280 | 1.7568 | 0.52 | 0.6488 | 2.5294 | 0.52 | 0.5132 | 0.2208 | 0.2442 |
No log | 41.0 | 287 | 1.6896 | 0.5275 | 0.6362 | 2.5489 | 0.5275 | 0.5141 | 0.2045 | 0.2323 |
No log | 42.0 | 294 | 1.7193 | 0.5275 | 0.6517 | 2.5525 | 0.5275 | 0.5232 | 0.1986 | 0.2467 |
No log | 43.0 | 301 | 1.7199 | 0.535 | 0.6403 | 2.5974 | 0.535 | 0.5279 | 0.2104 | 0.2432 |
No log | 44.0 | 308 | 1.6594 | 0.5375 | 0.6330 | 2.4854 | 0.5375 | 0.5316 | 0.2015 | 0.2321 |
No log | 45.0 | 315 | 1.6543 | 0.5275 | 0.6239 | 2.4955 | 0.5275 | 0.5223 | 0.2144 | 0.2308 |
No log | 46.0 | 322 | 1.6490 | 0.5425 | 0.6262 | 2.5215 | 0.5425 | 0.5358 | 0.2104 | 0.2273 |
No log | 47.0 | 329 | 1.6570 | 0.54 | 0.6233 | 2.5454 | 0.54 | 0.5380 | 0.2047 | 0.2301 |
No log | 48.0 | 336 | 1.6359 | 0.5375 | 0.6218 | 2.5546 | 0.5375 | 0.5320 | 0.2171 | 0.2257 |
No log | 49.0 | 343 | 1.6320 | 0.55 | 0.6214 | 2.4958 | 0.55 | 0.5452 | 0.2014 | 0.2267 |
No log | 50.0 | 350 | 1.6230 | 0.53 | 0.6208 | 2.4979 | 0.53 | 0.5243 | 0.2017 | 0.2315 |
No log | 51.0 | 357 | 1.6374 | 0.535 | 0.6257 | 2.4644 | 0.535 | 0.5293 | 0.2038 | 0.2286 |
No log | 52.0 | 364 | 1.6190 | 0.5375 | 0.6199 | 2.5279 | 0.5375 | 0.5310 | 0.1855 | 0.2290 |
No log | 53.0 | 371 | 1.6155 | 0.5475 | 0.6158 | 2.4738 | 0.5475 | 0.5435 | 0.1913 | 0.2239 |
No log | 54.0 | 378 | 1.6131 | 0.5425 | 0.6184 | 2.4982 | 0.5425 | 0.5377 | 0.1969 | 0.2248 |
No log | 55.0 | 385 | 1.6035 | 0.545 | 0.6138 | 2.4690 | 0.545 | 0.5406 | 0.2164 | 0.2223 |
No log | 56.0 | 392 | 1.5990 | 0.54 | 0.6153 | 2.4701 | 0.54 | 0.5356 | 0.2019 | 0.2249 |
No log | 57.0 | 399 | 1.6024 | 0.5425 | 0.6153 | 2.4626 | 0.5425 | 0.5375 | 0.1826 | 0.2237 |
No log | 58.0 | 406 | 1.5935 | 0.545 | 0.6141 | 2.4390 | 0.545 | 0.5415 | 0.1933 | 0.2238 |
No log | 59.0 | 413 | 1.6016 | 0.545 | 0.6137 | 2.4640 | 0.545 | 0.5401 | 0.2021 | 0.2230 |
No log | 60.0 | 420 | 1.5976 | 0.54 | 0.6146 | 2.4618 | 0.54 | 0.5355 | 0.1912 | 0.2245 |
No log | 61.0 | 427 | 1.5984 | 0.545 | 0.6133 | 2.4683 | 0.545 | 0.5408 | 0.1971 | 0.2228 |
No log | 62.0 | 434 | 1.5941 | 0.54 | 0.6131 | 2.4639 | 0.54 | 0.5358 | 0.1898 | 0.2236 |
No log | 63.0 | 441 | 1.5936 | 0.545 | 0.6123 | 2.4689 | 0.545 | 0.5404 | 0.1953 | 0.2222 |
No log | 64.0 | 448 | 1.5970 | 0.5425 | 0.6138 | 2.4647 | 0.5425 | 0.5384 | 0.2015 | 0.2238 |
No log | 65.0 | 455 | 1.5943 | 0.545 | 0.6130 | 2.4963 | 0.545 | 0.5400 | 0.1979 | 0.2229 |
No log | 66.0 | 462 | 1.5936 | 0.545 | 0.6127 | 2.4977 | 0.545 | 0.5400 | 0.1933 | 0.2229 |
No log | 67.0 | 469 | 1.5928 | 0.5425 | 0.6127 | 2.4965 | 0.5425 | 0.5381 | 0.1976 | 0.2233 |
No log | 68.0 | 476 | 1.5946 | 0.5425 | 0.6128 | 2.4768 | 0.5425 | 0.5383 | 0.2149 | 0.2233 |
No log | 69.0 | 483 | 1.5924 | 0.54 | 0.6126 | 2.4946 | 0.54 | 0.5356 | 0.2094 | 0.2233 |
No log | 70.0 | 490 | 1.5921 | 0.54 | 0.6120 | 2.4964 | 0.54 | 0.5356 | 0.1801 | 0.2230 |
No log | 71.0 | 497 | 1.5926 | 0.54 | 0.6126 | 2.4955 | 0.54 | 0.5356 | 0.2039 | 0.2235 |
0.3138 | 72.0 | 504 | 1.5916 | 0.5425 | 0.6121 | 2.4964 | 0.5425 | 0.5366 | 0.1898 | 0.2229 |
0.3138 | 73.0 | 511 | 1.5917 | 0.54 | 0.6119 | 2.4966 | 0.54 | 0.5356 | 0.2039 | 0.2231 |
0.3138 | 74.0 | 518 | 1.5918 | 0.54 | 0.6123 | 2.4964 | 0.54 | 0.5351 | 0.2035 | 0.2229 |
0.3138 | 75.0 | 525 | 1.5912 | 0.54 | 0.6118 | 2.4975 | 0.54 | 0.5351 | 0.2059 | 0.2228 |
0.3138 | 76.0 | 532 | 1.5918 | 0.54 | 0.6124 | 2.4965 | 0.54 | 0.5351 | 0.1971 | 0.2231 |
0.3138 | 77.0 | 539 | 1.5919 | 0.5425 | 0.6120 | 2.4974 | 0.5425 | 0.5358 | 0.2087 | 0.2227 |
0.3138 | 78.0 | 546 | 1.5903 | 0.54 | 0.6118 | 2.4978 | 0.54 | 0.5341 | 0.2169 | 0.2228 |
0.3138 | 79.0 | 553 | 1.5922 | 0.54 | 0.6124 | 2.4976 | 0.54 | 0.5351 | 0.2109 | 0.2234 |
0.3138 | 80.0 | 560 | 1.5914 | 0.54 | 0.6122 | 2.4983 | 0.54 | 0.5345 | 0.2041 | 0.2228 |
0.3138 | 81.0 | 567 | 1.5907 | 0.54 | 0.6119 | 2.4981 | 0.54 | 0.5345 | 0.2128 | 0.2226 |
0.3138 | 82.0 | 574 | 1.5921 | 0.5425 | 0.6124 | 2.4986 | 0.5425 | 0.5362 | 0.2084 | 0.2227 |
0.3138 | 83.0 | 581 | 1.5918 | 0.5425 | 0.6125 | 2.4987 | 0.5425 | 0.5362 | 0.2038 | 0.2230 |
0.3138 | 84.0 | 588 | 1.5902 | 0.54 | 0.6120 | 2.4989 | 0.54 | 0.5345 | 0.2043 | 0.2226 |
0.3138 | 85.0 | 595 | 1.5919 | 0.5425 | 0.6124 | 2.4988 | 0.5425 | 0.5360 | 0.1998 | 0.2228 |
0.3138 | 86.0 | 602 | 1.5916 | 0.5425 | 0.6124 | 2.4990 | 0.5425 | 0.5362 | 0.2079 | 0.2227 |
0.3138 | 87.0 | 609 | 1.5906 | 0.54 | 0.6120 | 2.4990 | 0.54 | 0.5345 | 0.2037 | 0.2227 |
0.3138 | 88.0 | 616 | 1.5908 | 0.54 | 0.6120 | 2.4989 | 0.54 | 0.5345 | 0.2091 | 0.2230 |
0.3138 | 89.0 | 623 | 1.5909 | 0.54 | 0.6120 | 2.4995 | 0.54 | 0.5344 | 0.2113 | 0.2228 |
0.3138 | 90.0 | 630 | 1.5906 | 0.54 | 0.6119 | 2.4996 | 0.54 | 0.5345 | 0.1969 | 0.2228 |
0.3138 | 91.0 | 637 | 1.5911 | 0.5425 | 0.6121 | 2.4999 | 0.5425 | 0.5360 | 0.1954 | 0.2226 |
0.3138 | 92.0 | 644 | 1.5909 | 0.54 | 0.6121 | 2.4994 | 0.54 | 0.5344 | 0.1928 | 0.2228 |
0.3138 | 93.0 | 651 | 1.5907 | 0.5425 | 0.6121 | 2.4999 | 0.5425 | 0.5360 | 0.2034 | 0.2225 |
0.3138 | 94.0 | 658 | 1.5910 | 0.5425 | 0.6122 | 2.4996 | 0.5425 | 0.5360 | 0.1974 | 0.2227 |
0.3138 | 95.0 | 665 | 1.5909 | 0.5375 | 0.6121 | 2.4995 | 0.5375 | 0.5319 | 0.1990 | 0.2230 |
0.3138 | 96.0 | 672 | 1.5907 | 0.5375 | 0.6120 | 2.4997 | 0.5375 | 0.5318 | 0.1980 | 0.2229 |
0.3138 | 97.0 | 679 | 1.5907 | 0.54 | 0.6120 | 2.4998 | 0.54 | 0.5344 | 0.1900 | 0.2228 |
0.3138 | 98.0 | 686 | 1.5907 | 0.5425 | 0.6120 | 2.4999 | 0.5425 | 0.5362 | 0.1899 | 0.2226 |
0.3138 | 99.0 | 693 | 1.5908 | 0.54 | 0.6121 | 2.4999 | 0.54 | 0.5334 | 0.1936 | 0.2228 |
0.3138 | 100.0 | 700 | 1.5908 | 0.54 | 0.6121 | 2.4999 | 0.54 | 0.5334 | 0.1895 | 0.2228 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2