ldos's picture
End of training
a9e4cd4
metadata
license: apache-2.0
base_model: t5-small
tags:
  - generated_from_trainer
metrics:
  - rouge
model-index:
  - name: text_shortening_model_v30
    results: []

text_shortening_model_v30

This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 2.6784
  • Rouge1: 0.4871
  • Rouge2: 0.2579
  • Rougel: 0.428
  • Rougelsum: 0.4272
  • Bert precision: 0.8743
  • Bert recall: 0.8706
  • Average word count: 8.4775
  • Max word count: 17
  • Min word count: 3
  • Average token count: 12.9249
  • % shortened texts with length > 12: 9.3093

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: 0.0003
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 100

Training results

Training Loss Epoch Step Validation Loss Rouge1 Rouge2 Rougel Rougelsum Bert precision Bert recall Average word count Max word count Min word count Average token count % shortened texts with length > 12
1.2044 1.0 145 1.6064 0.5052 0.2865 0.4472 0.448 0.8751 0.8756 8.8979 17 3 13.4024 12.6126
1.0041 2.0 290 1.4900 0.5154 0.2921 0.4554 0.4542 0.8735 0.878 9.3724 17 3 13.8529 17.7177
0.8935 3.0 435 1.4617 0.5181 0.2968 0.4607 0.4622 0.8751 0.8818 9.4024 16 4 14.1171 17.1171
0.8028 4.0 580 1.4744 0.5103 0.2966 0.4497 0.4496 0.8797 0.8725 8.1982 17 4 12.5706 8.1081
0.7395 5.0 725 1.4797 0.5121 0.3016 0.4548 0.4554 0.8796 0.8761 8.4985 16 3 12.985 10.8108
0.6986 6.0 870 1.5154 0.5218 0.2987 0.4554 0.4542 0.8808 0.879 8.7297 16 4 13.0691 14.1141
0.6527 7.0 1015 1.5347 0.5083 0.2876 0.4494 0.4485 0.8797 0.8763 8.5526 16 4 13.012 11.4114
0.588 8.0 1160 1.5578 0.4984 0.2752 0.4403 0.4399 0.8746 0.8728 8.6336 17 4 13.006 10.8108
0.5705 9.0 1305 1.6569 0.5152 0.2902 0.4544 0.454 0.8803 0.8764 8.5135 16 4 13.1592 9.9099
0.5601 10.0 1450 1.6651 0.5246 0.2837 0.4572 0.4579 0.8777 0.8807 8.979 16 4 13.6607 12.012
0.523 11.0 1595 1.7085 0.5149 0.2854 0.4508 0.4507 0.879 0.8789 8.7718 17 4 13.2613 10.8108
0.5032 12.0 1740 1.7886 0.5107 0.2817 0.4457 0.4457 0.8778 0.8772 8.8378 17 4 13.4204 11.7117
0.4872 13.0 1885 1.8073 0.5097 0.2808 0.4439 0.4441 0.8786 0.8758 8.6306 16 4 13.1562 9.6096
0.4703 14.0 2030 1.8436 0.5059 0.2754 0.4456 0.4457 0.8769 0.8756 8.6817 17 4 13.1471 9.9099
0.4598 15.0 2175 1.9150 0.5148 0.2794 0.4532 0.4532 0.8798 0.8775 8.6907 18 4 13.1021 11.4114
0.4385 16.0 2320 1.9319 0.4966 0.2666 0.4402 0.4406 0.8771 0.8724 8.2703 16 4 12.7237 7.8078
0.4306 17.0 2465 1.9821 0.5041 0.2763 0.4449 0.4448 0.8788 0.8752 8.5105 16 4 13.0541 9.3093
0.4154 18.0 2610 2.0345 0.5066 0.2746 0.4467 0.4461 0.8796 0.8732 8.1922 16 3 12.6186 7.8078
0.3995 19.0 2755 2.0671 0.4954 0.2707 0.4411 0.4416 0.8773 0.8721 8.4505 17 4 12.8468 8.7087
0.4053 20.0 2900 2.1265 0.4975 0.2704 0.4365 0.4364 0.8767 0.873 8.5075 17 3 13.0571 9.009
0.3812 21.0 3045 2.2077 0.5011 0.2733 0.4406 0.4411 0.8764 0.8756 8.7958 17 3 13.4084 12.012
0.3856 22.0 3190 2.2043 0.4956 0.2603 0.4358 0.4361 0.8775 0.8729 8.2913 17 3 12.8078 8.7087
0.3805 23.0 3335 2.2201 0.5015 0.2698 0.4421 0.4427 0.8789 0.8728 8.2402 17 3 12.5856 8.1081
0.3741 24.0 3480 2.2269 0.5029 0.2652 0.4412 0.4413 0.8767 0.8743 8.5856 16 4 13.039 10.2102
0.3697 25.0 3625 2.2596 0.4956 0.2674 0.436 0.4359 0.8765 0.8728 8.4895 17 4 12.9129 9.9099
0.3663 26.0 3770 2.2506 0.4891 0.2572 0.432 0.432 0.8749 0.8716 8.4865 17 4 12.8498 6.9069
0.3409 27.0 3915 2.2893 0.4958 0.2635 0.4328 0.4327 0.8772 0.8727 8.3994 17 3 12.8228 9.6096
0.3524 28.0 4060 2.3127 0.4907 0.2597 0.4322 0.4329 0.8751 0.8712 8.4084 16 4 12.7718 8.1081
0.3379 29.0 4205 2.3167 0.4958 0.2674 0.4374 0.4368 0.8772 0.8737 8.4234 16 4 12.8138 7.2072
0.3472 30.0 4350 2.3157 0.4987 0.2713 0.4415 0.4403 0.8788 0.8736 8.3634 17 3 12.6517 7.2072
0.3353 31.0 4495 2.3506 0.4991 0.2631 0.4375 0.436 0.8764 0.8744 8.6396 17 4 13.1502 9.6096
0.3466 32.0 4640 2.3594 0.4897 0.2593 0.4307 0.4301 0.8777 0.8711 8.1712 16 4 12.6126 5.4054
0.3406 33.0 4785 2.3632 0.495 0.2746 0.4401 0.4397 0.8772 0.8732 8.5556 16 4 13.027 8.4084
0.3382 34.0 4930 2.3505 0.4856 0.261 0.4306 0.4295 0.8758 0.8693 8.2733 17 3 12.6366 7.5075
0.3392 35.0 5075 2.3665 0.4972 0.2719 0.4376 0.4372 0.8764 0.8741 8.6847 17 4 13.1532 9.3093
0.3465 36.0 5220 2.3837 0.4981 0.2722 0.441 0.4411 0.876 0.8738 8.6607 17 4 13.1982 12.3123
0.3377 37.0 5365 2.3984 0.4832 0.2623 0.4294 0.4285 0.8737 0.8697 8.5225 17 4 12.9399 10.5105
0.3523 38.0 5510 2.3843 0.495 0.2671 0.438 0.4368 0.8754 0.873 8.5886 17 3 13.1111 7.2072
0.3261 39.0 5655 2.4337 0.4948 0.2666 0.4378 0.4369 0.8771 0.8726 8.4655 17 4 12.8919 9.009
0.3262 40.0 5800 2.4149 0.4971 0.2691 0.438 0.4375 0.8772 0.8717 8.4505 16 4 12.9249 8.1081
0.3307 41.0 5945 2.4352 0.4834 0.2585 0.4261 0.4256 0.8746 0.8697 8.4024 17 3 12.8859 9.6096
0.3226 42.0 6090 2.4241 0.488 0.2584 0.4318 0.4315 0.8756 0.8706 8.4444 17 3 12.8288 8.7087
0.34 43.0 6235 2.4485 0.4891 0.2589 0.4326 0.432 0.8758 0.8705 8.3243 17 4 12.7898 6.6066
0.3425 44.0 6380 2.4457 0.4865 0.26 0.4293 0.4287 0.8733 0.8713 8.6336 16 3 13.1922 9.6096
0.3201 45.0 6525 2.4535 0.4811 0.2473 0.4243 0.4237 0.8751 0.8697 8.3093 17 3 12.7748 8.4084
0.3094 46.0 6670 2.4918 0.4916 0.2614 0.4351 0.4342 0.8758 0.8726 8.5706 17 3 13.039 10.2102
0.3262 47.0 6815 2.4839 0.4822 0.255 0.425 0.4237 0.8719 0.869 8.5375 17 4 12.976 9.009
0.3186 48.0 6960 2.4966 0.486 0.2492 0.4276 0.4264 0.8738 0.8707 8.4745 17 3 12.955 6.6066
0.3231 49.0 7105 2.4978 0.4889 0.2661 0.4343 0.434 0.8767 0.871 8.4505 17 3 12.8468 9.009
0.3294 50.0 7250 2.4731 0.4916 0.2683 0.4374 0.4373 0.877 0.8726 8.4955 17 4 12.9369 9.3093
0.3172 51.0 7395 2.4922 0.4861 0.2573 0.4314 0.431 0.8759 0.87 8.3003 17 4 12.6907 7.8078
0.3247 52.0 7540 2.5044 0.4802 0.2495 0.4281 0.4282 0.8737 0.8698 8.4715 17 4 12.9009 8.1081
0.3132 53.0 7685 2.5168 0.4832 0.2558 0.4273 0.4268 0.8736 0.8703 8.5706 17 3 12.967 9.3093
0.3285 54.0 7830 2.5296 0.4882 0.26 0.4323 0.4319 0.8754 0.8724 8.5495 17 3 13.0541 8.7087
0.3111 55.0 7975 2.5529 0.4829 0.2561 0.4268 0.4262 0.874 0.8694 8.4474 17 3 12.9339 7.2072
0.3194 56.0 8120 2.5903 0.49 0.2614 0.4337 0.4329 0.8747 0.8719 8.5946 17 3 13.0931 8.1081
0.3144 57.0 8265 2.5787 0.4859 0.2593 0.4315 0.4303 0.8739 0.8698 8.5195 17 4 12.8679 8.4084
0.2972 58.0 8410 2.5759 0.4848 0.2565 0.4291 0.4279 0.8738 0.8697 8.5165 17 3 12.9219 8.1081
0.3209 59.0 8555 2.5609 0.4792 0.246 0.4212 0.4201 0.8723 0.8678 8.4114 17 3 12.8799 6.9069
0.3148 60.0 8700 2.5758 0.481 0.2454 0.4243 0.4231 0.874 0.8688 8.3664 16 3 12.7628 7.5075
0.3026 61.0 8845 2.5819 0.4804 0.2555 0.4231 0.4231 0.8738 0.8689 8.4204 17 3 12.7628 8.4084
0.3074 62.0 8990 2.5882 0.4893 0.2627 0.431 0.4303 0.8753 0.8715 8.4895 17 3 12.8889 8.7087
0.3013 63.0 9135 2.5865 0.4835 0.2599 0.426 0.4251 0.8743 0.8707 8.4865 17 4 12.964 8.7087
0.3274 64.0 9280 2.5957 0.4928 0.2649 0.436 0.4353 0.8738 0.8734 8.8018 17 3 13.2823 11.4114
0.2928 65.0 9425 2.5846 0.4888 0.2653 0.4365 0.4356 0.8763 0.8713 8.2973 17 3 12.6637 8.1081
0.3261 66.0 9570 2.5704 0.4901 0.267 0.4386 0.4374 0.8759 0.871 8.3303 17 4 12.7838 6.6066
0.3153 67.0 9715 2.6023 0.4897 0.2611 0.4311 0.4301 0.8749 0.872 8.6426 17 3 13.0691 10.8108
0.3185 68.0 9860 2.5831 0.4862 0.2579 0.4257 0.4247 0.8735 0.8718 8.6486 17 4 13.1441 12.012
0.3054 69.0 10005 2.5949 0.4831 0.2575 0.4247 0.4239 0.8728 0.87 8.5405 17 4 13.036 9.9099
0.3006 70.0 10150 2.5822 0.4853 0.252 0.4255 0.4243 0.8735 0.87 8.5495 17 3 13.0 10.5105
0.3092 71.0 10295 2.5743 0.4903 0.2595 0.432 0.4315 0.8759 0.8719 8.4474 17 3 12.8559 8.7087
0.2928 72.0 10440 2.5905 0.4918 0.2665 0.4356 0.4347 0.876 0.8724 8.4474 17 4 12.8679 8.4084
0.3021 73.0 10585 2.6171 0.4957 0.266 0.4368 0.4354 0.8764 0.873 8.5676 17 3 12.964 11.1111
0.3047 74.0 10730 2.6233 0.492 0.2655 0.4341 0.4328 0.8753 0.8715 8.5736 17 3 12.952 10.5105
0.3043 75.0 10875 2.6405 0.4887 0.2623 0.4318 0.4309 0.8756 0.8704 8.4895 17 3 12.8679 9.9099
0.305 76.0 11020 2.6171 0.4942 0.2687 0.4381 0.4372 0.8766 0.8724 8.5586 17 3 12.9369 10.8108
0.3127 77.0 11165 2.6289 0.4959 0.2646 0.4366 0.4357 0.8767 0.8731 8.5766 17 3 13.006 12.012
0.2945 78.0 11310 2.6453 0.4881 0.2589 0.4272 0.4261 0.8753 0.8711 8.5375 17 3 12.8739 9.3093
0.2844 79.0 11455 2.6543 0.4895 0.2565 0.4294 0.4288 0.8753 0.8718 8.5616 17 3 12.997 11.7117
0.3188 80.0 11600 2.6556 0.4919 0.2677 0.4328 0.4318 0.8756 0.8712 8.5345 17 3 12.973 9.9099
0.2857 81.0 11745 2.6696 0.4914 0.2666 0.434 0.4332 0.8761 0.8717 8.4595 17 3 12.8829 10.5105
0.3091 82.0 11890 2.6577 0.4986 0.2718 0.4397 0.4388 0.8766 0.8741 8.6276 17 3 13.1441 10.8108
0.3115 83.0 12035 2.6720 0.4944 0.266 0.4364 0.4351 0.8766 0.8725 8.4925 17 3 12.9309 9.3093
0.2947 84.0 12180 2.6490 0.4955 0.2628 0.4347 0.4343 0.8767 0.873 8.4985 17 3 13.018 7.5075
0.312 85.0 12325 2.6425 0.4928 0.2689 0.4364 0.4358 0.8763 0.8728 8.5766 17 3 13.0631 9.9099
0.3081 86.0 12470 2.6314 0.4904 0.2648 0.4327 0.432 0.875 0.8722 8.6246 17 3 13.1411 10.5105
0.3043 87.0 12615 2.6485 0.4863 0.259 0.4273 0.4259 0.8736 0.8709 8.5736 17 3 13.0901 9.6096
0.3034 88.0 12760 2.6402 0.4867 0.2604 0.4279 0.4274 0.8739 0.871 8.5706 17 3 13.0751 8.1081
0.3058 89.0 12905 2.6573 0.4926 0.2638 0.4348 0.4339 0.8762 0.872 8.4805 17 3 12.955 7.8078
0.2909 90.0 13050 2.6654 0.4955 0.2679 0.4357 0.4342 0.8756 0.8729 8.6817 17 3 13.1802 10.2102
0.3082 91.0 13195 2.6757 0.4942 0.2671 0.4362 0.4349 0.8756 0.8724 8.5796 17 3 13.0721 9.6096
0.3016 92.0 13340 2.6791 0.4933 0.2657 0.4351 0.4345 0.875 0.8722 8.6336 17 3 13.1441 9.9099
0.2993 93.0 13485 2.6814 0.493 0.2658 0.433 0.4318 0.8747 0.8726 8.6997 17 3 13.2462 11.1111
0.3022 94.0 13630 2.6698 0.4929 0.2638 0.4334 0.4324 0.8751 0.8723 8.5976 17 3 13.0961 9.3093
0.2921 95.0 13775 2.6665 0.4867 0.2586 0.4294 0.4284 0.8744 0.8709 8.4955 17 3 12.988 8.4084
0.3034 96.0 13920 2.6704 0.4854 0.2574 0.4275 0.4266 0.8742 0.8704 8.4805 17 3 12.9429 8.7087
0.3063 97.0 14065 2.6749 0.4863 0.2576 0.4275 0.4266 0.8743 0.8707 8.4805 17 3 12.9369 8.7087
0.2984 98.0 14210 2.6772 0.4858 0.258 0.4274 0.4264 0.8739 0.8704 8.5105 17 3 12.97 9.6096
0.2942 99.0 14355 2.6784 0.4872 0.2595 0.4279 0.427 0.874 0.8704 8.5075 17 3 12.967 9.6096
0.2866 100.0 14500 2.6784 0.4871 0.2579 0.428 0.4272 0.8743 0.8706 8.4775 17 3 12.9249 9.3093

Framework versions

  • Transformers 4.33.1
  • Pytorch 2.0.1+cu118
  • Datasets 2.14.5
  • Tokenizers 0.13.3