ldos's picture
End of training
26a5405
metadata
license: apache-2.0
base_model: t5-small
tags:
  - generated_from_trainer
metrics:
  - rouge
model-index:
  - name: text_shortening_model_v9
    results: []

text_shortening_model_v9

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

  • Loss: 1.7285
  • Rouge1: 0.5919
  • Rouge2: 0.3742
  • Rougel: 0.5529
  • Rougelsum: 0.5532
  • Bert precision: 0.8979
  • Bert recall: 0.9029
  • Average word count: 11.1929
  • Max word count: 17
  • Min word count: 7
  • Average token count: 16.3286
  • % shortened texts with length > 12: 22.1429

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: 32
  • eval_batch_size: 32
  • 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
1.2656 1.0 16 1.6819 0.5512 0.3185 0.4947 0.4946 0.8804 0.8891 11.8643 18 5 17.0071 45.7143
1.1187 2.0 32 1.5924 0.567 0.3403 0.5157 0.5151 0.8857 0.8954 11.8214 18 3 16.7786 45.7143
1.0753 3.0 48 1.5304 0.5832 0.3555 0.5319 0.5304 0.8881 0.8998 11.9571 18 4 17.0357 46.4286
1.0235 4.0 64 1.4952 0.5785 0.3453 0.5277 0.527 0.8875 0.9003 11.8857 17 6 17.0286 42.8571
0.9861 5.0 80 1.4627 0.5894 0.3606 0.5388 0.5379 0.8885 0.901 11.9429 17 6 17.1929 43.5714
0.9616 6.0 96 1.4499 0.59 0.3567 0.536 0.5355 0.8877 0.9019 12.0071 18 6 17.2714 42.8571
0.9193 7.0 112 1.4335 0.5912 0.3627 0.5427 0.5419 0.8877 0.9025 11.9786 17 6 17.3571 40.7143
0.8959 8.0 128 1.4193 0.5866 0.3583 0.5346 0.5337 0.8887 0.9016 11.7714 17 6 17.1143 38.5714
0.8834 9.0 144 1.4090 0.5979 0.369 0.5469 0.5464 0.8908 0.9042 11.7 16 6 17.2071 37.8571
0.8468 10.0 160 1.4035 0.5977 0.3678 0.5473 0.5469 0.8916 0.9048 11.7643 17 6 17.2071 35.7143
0.8297 11.0 176 1.3956 0.5986 0.365 0.549 0.5475 0.8934 0.9046 11.5857 16 6 16.9429 32.8571
0.8275 12.0 192 1.3934 0.6027 0.3731 0.555 0.5551 0.8934 0.9049 11.6143 17 6 16.9286 32.8571
0.8072 13.0 208 1.3915 0.5973 0.3672 0.5484 0.5472 0.8921 0.905 11.7214 16 6 17.0857 35.7143
0.7744 14.0 224 1.3972 0.6006 0.3707 0.5544 0.5529 0.8947 0.9051 11.5214 16 6 16.8714 33.5714
0.7626 15.0 240 1.3910 0.6039 0.3745 0.5586 0.5576 0.8962 0.9053 11.5071 16 6 16.7071 36.4286
0.7564 16.0 256 1.3918 0.6046 0.3739 0.5571 0.5563 0.8943 0.906 11.7286 17 6 17.0214 40.0
0.7599 17.0 272 1.3822 0.6025 0.3753 0.5549 0.5542 0.8939 0.9059 11.6571 16 6 17.0429 35.7143
0.7331 18.0 288 1.3885 0.6019 0.3705 0.5548 0.5539 0.8935 0.9048 11.65 16 6 17.0357 34.2857
0.7227 19.0 304 1.3916 0.6084 0.3825 0.563 0.5628 0.8991 0.9069 11.2214 16 6 16.5786 27.1429
0.6906 20.0 320 1.4023 0.6065 0.3797 0.5579 0.5579 0.8934 0.9067 11.7714 16 7 17.1357 37.1429
0.6917 21.0 336 1.4052 0.6095 0.3831 0.5621 0.5623 0.8965 0.9072 11.4357 16 6 16.7786 31.4286
0.6867 22.0 352 1.4104 0.6026 0.3807 0.5558 0.5561 0.8928 0.9057 11.5857 16 6 17.0643 31.4286
0.6995 23.0 368 1.4127 0.5999 0.3744 0.5514 0.5511 0.8941 0.9034 11.3571 16 6 16.6714 29.2857
0.6699 24.0 384 1.4217 0.6003 0.3804 0.5558 0.5551 0.8945 0.906 11.4714 16 7 16.8857 29.2857
0.6598 25.0 400 1.4344 0.5975 0.3744 0.552 0.5517 0.8943 0.9053 11.4429 16 6 16.7857 29.2857
0.6592 26.0 416 1.4340 0.6081 0.3868 0.5617 0.5614 0.8964 0.9071 11.3786 16 7 16.8 27.8571
0.6651 27.0 432 1.4375 0.6005 0.3741 0.553 0.553 0.8947 0.9042 11.3714 16 6 16.7071 28.5714
0.6409 28.0 448 1.4511 0.5977 0.3713 0.5508 0.5508 0.8959 0.9033 11.05 16 6 16.45 22.1429
0.6373 29.0 464 1.4670 0.5918 0.3655 0.5426 0.5426 0.8933 0.9026 11.3429 16 7 16.8071 25.7143
0.6284 30.0 480 1.4591 0.5973 0.3782 0.5497 0.5498 0.8947 0.904 11.3 16 7 16.8 24.2857
0.6214 31.0 496 1.4709 0.5987 0.3806 0.5543 0.5543 0.8963 0.9041 11.2214 16 6 16.6714 25.7143
0.6086 32.0 512 1.4839 0.5874 0.3667 0.5442 0.5434 0.8942 0.9016 11.1357 16 6 16.5429 26.4286
0.6102 33.0 528 1.4852 0.5928 0.3746 0.5479 0.5474 0.8954 0.9022 11.1286 16 6 16.5071 24.2857
0.6118 34.0 544 1.4869 0.5962 0.3766 0.5488 0.5486 0.8948 0.9035 11.4 16 7 16.7643 27.1429
0.605 35.0 560 1.4881 0.5943 0.3746 0.5461 0.5457 0.8942 0.9019 11.3143 16 7 16.7929 26.4286
0.6039 36.0 576 1.4854 0.5903 0.3716 0.5431 0.5431 0.8957 0.9014 11.1 16 7 16.45 24.2857
0.5777 37.0 592 1.4901 0.5922 0.3685 0.5461 0.546 0.8943 0.9042 11.3786 16 7 16.8143 26.4286
0.5634 38.0 608 1.4975 0.594 0.3721 0.5454 0.5446 0.8958 0.9019 11.0929 16 7 16.4286 22.8571
0.5794 39.0 624 1.5088 0.5963 0.3792 0.5515 0.5508 0.896 0.9026 11.2429 16 7 16.55 24.2857
0.5825 40.0 640 1.5150 0.5951 0.3736 0.5512 0.5502 0.895 0.9031 11.3786 16 6 16.6643 27.8571
0.5632 41.0 656 1.5230 0.5998 0.3731 0.5571 0.5561 0.9 0.9037 11.0714 16 6 16.1214 22.1429
0.5544 42.0 672 1.5356 0.6036 0.3798 0.5628 0.5628 0.8987 0.9046 11.2143 16 7 16.3143 22.8571
0.5672 43.0 688 1.5493 0.5944 0.3671 0.5502 0.5504 0.8954 0.9024 11.3786 16 7 16.6 25.0
0.551 44.0 704 1.5563 0.5859 0.362 0.543 0.5426 0.8957 0.9002 11.1214 15 7 16.35 23.5714
0.543 45.0 720 1.5601 0.592 0.3643 0.5484 0.5481 0.8968 0.9014 11.0929 17 7 16.3 22.8571
0.5352 46.0 736 1.5680 0.6039 0.3783 0.5618 0.5614 0.8987 0.905 11.1929 17 7 16.4071 23.5714
0.528 47.0 752 1.5732 0.595 0.3721 0.5562 0.5558 0.8968 0.9024 11.1643 17 7 16.3714 25.0
0.528 48.0 768 1.5749 0.5933 0.372 0.5539 0.5537 0.896 0.9026 11.2643 17 7 16.4857 25.7143
0.5296 49.0 784 1.5795 0.596 0.3726 0.554 0.5541 0.897 0.904 11.2571 17 7 16.4571 26.4286
0.5235 50.0 800 1.5828 0.5916 0.3679 0.5484 0.548 0.8951 0.9019 11.2643 17 7 16.4571 27.1429
0.5168 51.0 816 1.5879 0.5917 0.368 0.5473 0.5465 0.8962 0.9006 11.1857 17 7 16.2286 25.7143
0.5133 52.0 832 1.5932 0.5928 0.3665 0.5473 0.5465 0.8979 0.9018 11.1643 17 7 16.2643 21.4286
0.5036 53.0 848 1.6016 0.5927 0.3703 0.5508 0.5511 0.8949 0.9012 11.3286 17 7 16.4143 26.4286
0.492 54.0 864 1.6074 0.5922 0.37 0.5496 0.5493 0.8953 0.9021 11.3643 17 7 16.5214 26.4286
0.5184 55.0 880 1.6153 0.5953 0.3714 0.5542 0.5536 0.8963 0.9027 11.3 17 7 16.5 24.2857
0.5057 56.0 896 1.6311 0.5874 0.3636 0.5424 0.5425 0.896 0.9009 11.0857 17 7 16.2429 21.4286
0.5053 57.0 912 1.6356 0.5835 0.3623 0.5411 0.5408 0.8953 0.8996 11.1929 17 7 16.3143 25.7143
0.5016 58.0 928 1.6342 0.5908 0.3679 0.5475 0.5472 0.8966 0.9011 11.1214 17 7 16.2929 23.5714
0.4921 59.0 944 1.6312 0.5899 0.3719 0.5495 0.549 0.8966 0.9006 11.0429 17 7 16.1929 25.0
0.5051 60.0 960 1.6316 0.5989 0.3766 0.5572 0.5566 0.8964 0.9045 11.3214 17 7 16.6643 25.7143
0.4938 61.0 976 1.6377 0.6007 0.3812 0.5581 0.5578 0.898 0.903 11.1214 17 7 16.2357 25.0
0.4843 62.0 992 1.6437 0.5981 0.3844 0.5597 0.5595 0.8965 0.9033 11.1714 17 7 16.3286 26.4286
0.4894 63.0 1008 1.6473 0.594 0.3718 0.5525 0.5523 0.8951 0.903 11.2857 17 7 16.5071 28.5714
0.4956 64.0 1024 1.6549 0.5843 0.37 0.5449 0.5447 0.895 0.8995 11.0929 17 7 16.2 25.7143
0.4852 65.0 1040 1.6543 0.5947 0.3742 0.5553 0.555 0.8958 0.9024 11.35 17 7 16.55 27.8571
0.489 66.0 1056 1.6558 0.5922 0.3751 0.5546 0.5544 0.896 0.9014 11.1357 17 7 16.2857 25.7143
0.4852 67.0 1072 1.6619 0.591 0.376 0.5522 0.5523 0.8959 0.9016 11.1571 17 7 16.2571 23.5714
0.4847 68.0 1088 1.6699 0.5913 0.3781 0.556 0.5556 0.8969 0.901 11.0214 17 7 16.1357 22.8571
0.4685 69.0 1104 1.6720 0.5909 0.3755 0.5516 0.5517 0.8961 0.9015 11.2571 17 7 16.35 25.0
0.4799 70.0 1120 1.6734 0.586 0.3654 0.5448 0.5454 0.8937 0.8998 11.25 17 7 16.3214 24.2857
0.4781 71.0 1136 1.6765 0.5844 0.3634 0.5429 0.5428 0.8927 0.8996 11.35 17 7 16.4929 26.4286
0.4843 72.0 1152 1.6814 0.5864 0.3619 0.5426 0.5432 0.8928 0.9006 11.4286 17 7 16.5929 27.8571
0.4658 73.0 1168 1.6846 0.5888 0.3628 0.5431 0.5437 0.8941 0.9001 11.3214 17 7 16.4429 25.7143
0.4664 74.0 1184 1.6899 0.5885 0.3692 0.5473 0.5473 0.8949 0.9 11.1786 17 7 16.3143 22.1429
0.4805 75.0 1200 1.6954 0.5915 0.3765 0.5506 0.5511 0.8956 0.9013 11.2286 17 7 16.3643 23.5714
0.4708 76.0 1216 1.6964 0.5888 0.37 0.5479 0.5483 0.8964 0.9004 11.0571 17 7 16.1929 21.4286
0.4483 77.0 1232 1.6968 0.5881 0.3669 0.5455 0.5457 0.8954 0.8999 11.1214 17 7 16.2857 22.8571
0.4699 78.0 1248 1.6993 0.5908 0.369 0.5477 0.5481 0.8957 0.9015 11.1786 15 7 16.3857 24.2857
0.4657 79.0 1264 1.7014 0.5927 0.3734 0.5528 0.553 0.8971 0.9021 11.1429 15 7 16.3214 22.8571
0.4616 80.0 1280 1.7063 0.5919 0.3743 0.5531 0.5533 0.8975 0.9009 11.0714 15 7 16.25 20.7143
0.4706 81.0 1296 1.7087 0.5933 0.3728 0.5521 0.5525 0.8976 0.9015 11.0643 15 7 16.2429 21.4286
0.4557 82.0 1312 1.7109 0.5917 0.3717 0.5517 0.5515 0.8971 0.902 11.1429 17 7 16.35 22.8571
0.474 83.0 1328 1.7164 0.5918 0.3714 0.5507 0.5509 0.8967 0.9024 11.2357 17 7 16.4143 24.2857
0.4715 84.0 1344 1.7165 0.591 0.3717 0.5522 0.5533 0.8975 0.9025 11.1071 17 7 16.2857 22.8571
0.462 85.0 1360 1.7159 0.5892 0.3708 0.5479 0.5481 0.896 0.9021 11.2071 17 7 16.3714 23.5714
0.455 86.0 1376 1.7171 0.5943 0.379 0.5551 0.5559 0.898 0.9031 11.1929 17 7 16.3429 23.5714
0.4613 87.0 1392 1.7173 0.5894 0.371 0.5501 0.5507 0.8967 0.9018 11.2 17 7 16.3571 22.8571
0.4663 88.0 1408 1.7191 0.5895 0.3705 0.5505 0.5509 0.8968 0.9018 11.1857 17 7 16.3429 22.1429
0.4662 89.0 1424 1.7213 0.5893 0.3692 0.5498 0.5501 0.8961 0.9012 11.2214 17 7 16.3714 23.5714
0.4352 90.0 1440 1.7202 0.5886 0.3696 0.549 0.5498 0.8963 0.9015 11.2214 17 7 16.3714 23.5714
0.4567 91.0 1456 1.7193 0.5885 0.373 0.5509 0.5516 0.8968 0.9022 11.1929 17 7 16.3429 23.5714
0.4421 92.0 1472 1.7211 0.5885 0.3734 0.5498 0.5505 0.8962 0.9022 11.2429 17 7 16.3857 24.2857
0.4655 93.0 1488 1.7230 0.5925 0.3763 0.5537 0.5538 0.8977 0.9029 11.1929 17 7 16.35 23.5714
0.4431 94.0 1504 1.7246 0.5912 0.3765 0.5529 0.5531 0.898 0.903 11.1929 17 7 16.3286 22.8571
0.4493 95.0 1520 1.7258 0.5921 0.3756 0.5531 0.5535 0.8979 0.903 11.2357 17 7 16.3714 22.8571
0.4546 96.0 1536 1.7272 0.5918 0.375 0.5529 0.5533 0.8978 0.9029 11.2357 17 7 16.3643 23.5714
0.4558 97.0 1552 1.7279 0.5925 0.3744 0.5536 0.554 0.8979 0.9029 11.2071 17 7 16.3357 22.8571
0.4575 98.0 1568 1.7281 0.592 0.3746 0.5532 0.5533 0.8978 0.9029 11.2 17 7 16.3357 22.8571
0.441 99.0 1584 1.7285 0.5919 0.3742 0.5529 0.5532 0.8978 0.9029 11.1929 17 7 16.3286 22.1429
0.4529 100.0 1600 1.7285 0.5919 0.3742 0.5529 0.5532 0.8979 0.9029 11.1929 17 7 16.3286 22.1429

Framework versions

  • Transformers 4.32.1
  • Pytorch 2.0.1+cu118
  • Datasets 2.14.4
  • Tokenizers 0.13.3