Hugging Face's logo

ldos
/
text_shortening_model_v13

Text2Text Generation
Transformers
PyTorch
t5
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Community
1
text_shortening_model_v13 / README.md
ldos's picture
ldos
End of training
4067774
preview code
|
raw
history blame contribute delete
No virus
16 kB
metadata
license: apache-2.0
base_model: t5-small
tags:
  - generated_from_trainer
metrics:
  - rouge
model-index:
  - name: text_shortening_model_v13
    results: []

text_shortening_model_v13

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

  • Loss: 2.0548
  • Rouge1: 0.5772
  • Rouge2: 0.3353
  • Rougel: 0.5189
  • Rougelsum: 0.5189
  • Bert precision: 0.8941
  • Bert recall: 0.8987
  • Average word count: 11.2143
  • Max word count: 15
  • Min word count: 6
  • Average token count: 16.5071
  • % shortened texts with length > 12: 30.0

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: 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: 60

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
1.1544 1.0 62 1.6383 0.5496 0.3173 0.5011 0.5016 0.8821 0.8901 11.7214 17 4 17.0214 50.0
1.0203 2.0 124 1.5337 0.569 0.3214 0.5114 0.5112 0.8863 0.8968 11.9857 17 5 17.2214 50.7143
0.9474 3.0 186 1.5169 0.5754 0.3283 0.5196 0.5186 0.8874 0.8985 12.0857 17 6 17.2571 53.5714
0.8615 4.0 248 1.5058 0.5785 0.3368 0.5211 0.5198 0.8917 0.9006 11.7 17 6 16.8714 47.8571
0.8182 5.0 310 1.4855 0.5817 0.3284 0.5203 0.5195 0.8907 0.9 11.7357 17 6 16.9357 46.4286
0.784 6.0 372 1.4813 0.5862 0.3398 0.5242 0.5242 0.8918 0.9016 11.7 16 6 17.0 45.0
0.7749 7.0 434 1.4723 0.581 0.334 0.5241 0.5233 0.8951 0.8984 11.1929 16 6 16.3286 32.1429
0.7396 8.0 496 1.4936 0.5791 0.3402 0.5184 0.5183 0.8933 0.8992 11.4786 17 6 16.5571 34.2857
0.6856 9.0 558 1.5083 0.5757 0.3364 0.5174 0.5172 0.8944 0.8979 11.2 16 6 16.2357 30.7143
0.6679 10.0 620 1.5295 0.5814 0.3399 0.5271 0.5276 0.8915 0.9 11.7786 16 7 16.9143 40.0
0.6506 11.0 682 1.5363 0.5829 0.3491 0.5282 0.5283 0.8953 0.8994 11.3786 16 6 16.5286 33.5714
0.6521 12.0 744 1.5526 0.5645 0.3303 0.5095 0.5096 0.8914 0.8951 11.2286 16 5 16.4929 30.7143
0.6125 13.0 806 1.5787 0.5709 0.324 0.5097 0.5108 0.8906 0.8953 11.4214 16 6 16.6571 35.0
0.5915 14.0 868 1.5946 0.5757 0.3373 0.5152 0.5159 0.8926 0.8969 11.4071 16 6 16.5571 32.8571
0.5737 15.0 930 1.6204 0.577 0.3322 0.5219 0.5223 0.8918 0.8986 11.5929 16 6 16.8214 35.7143
0.5812 16.0 992 1.6372 0.5748 0.3243 0.52 0.5203 0.891 0.8977 11.6071 16 7 16.8214 37.8571
0.5468 17.0 1054 1.6514 0.5673 0.3304 0.5152 0.5152 0.895 0.8954 11.0 15 5 15.9929 26.4286
0.56 18.0 1116 1.6630 0.576 0.3273 0.5229 0.5228 0.8907 0.898 11.5786 16 6 16.8429 35.0
0.5548 19.0 1178 1.6868 0.5739 0.3262 0.5139 0.5135 0.8923 0.8972 11.3429 16 6 16.5929 33.5714
0.5338 20.0 1240 1.6954 0.5702 0.3295 0.518 0.5182 0.8914 0.8975 11.6 16 6 16.7429 37.8571
0.5323 21.0 1302 1.7255 0.585 0.3376 0.5262 0.5266 0.8938 0.9007 11.5643 16 6 16.7429 35.0
0.5075 22.0 1364 1.7320 0.5708 0.3272 0.5137 0.5144 0.8929 0.8953 11.3286 16 6 16.4143 32.1429
0.4916 23.0 1426 1.7601 0.5724 0.3276 0.5161 0.5171 0.8928 0.8965 11.3357 16 6 16.6 31.4286
0.4789 24.0 1488 1.7779 0.5726 0.3253 0.5128 0.513 0.8934 0.8964 11.4143 16 6 16.5 35.0
0.4851 25.0 1550 1.7970 0.575 0.3318 0.5204 0.521 0.8935 0.8982 11.4429 16 6 16.6714 31.4286
0.4682 26.0 1612 1.8094 0.5783 0.3376 0.5203 0.5213 0.8937 0.8984 11.4714 16 6 16.7571 30.0
0.4703 27.0 1674 1.8299 0.5814 0.3383 0.5208 0.5215 0.8934 0.8982 11.3929 16 6 16.6643 30.0
0.483 28.0 1736 1.8396 0.576 0.3394 0.5155 0.5162 0.8945 0.8975 11.3357 16 6 16.3857 28.5714
0.4712 29.0 1798 1.8567 0.5741 0.326 0.5125 0.5129 0.893 0.8981 11.5 16 6 16.5786 35.7143
0.4679 30.0 1860 1.8818 0.5855 0.3416 0.5239 0.5242 0.895 0.9005 11.4429 16 6 16.7143 33.5714
0.4653 31.0 1922 1.8758 0.5805 0.3378 0.5217 0.5222 0.894 0.8986 11.3357 16 6 16.4857 30.0
0.4484 32.0 1984 1.8920 0.5812 0.3363 0.5207 0.5206 0.8946 0.8991 11.3357 16 6 16.5143 30.0
0.4428 33.0 2046 1.8925 0.5832 0.3372 0.5195 0.5203 0.8968 0.8987 11.1286 16 6 16.2214 26.4286
0.4266 34.0 2108 1.9185 0.5736 0.3322 0.517 0.518 0.8952 0.8974 11.0214 15 6 16.1714 25.0
0.429 35.0 2170 1.9366 0.5829 0.3371 0.5224 0.5231 0.8965 0.8988 11.1643 16 6 16.25 27.1429
0.4034 36.0 2232 1.9510 0.5823 0.3392 0.5288 0.5288 0.8963 0.8986 11.1143 15 6 16.1214 30.0
0.4111 37.0 2294 1.9517 0.587 0.3426 0.529 0.5296 0.8959 0.9011 11.3857 16 6 16.55 31.4286
0.4318 38.0 2356 1.9450 0.5851 0.3444 0.5262 0.5268 0.8963 0.9009 11.2714 16 6 16.4571 30.0
0.4399 39.0 2418 1.9539 0.5772 0.3339 0.5164 0.5169 0.8958 0.8995 11.0929 15 6 16.2929 25.0
0.4268 40.0 2480 1.9620 0.5806 0.3319 0.5187 0.5188 0.8962 0.8983 11.0214 16 6 16.0643 26.4286
0.4119 41.0 2542 1.9939 0.5819 0.3408 0.5239 0.5238 0.8945 0.8992 11.3 16 6 16.4929 30.0
0.4061 42.0 2604 1.9714 0.5813 0.338 0.5214 0.5228 0.897 0.8997 11.05 16 6 16.2429 25.7143
0.4176 43.0 2666 1.9911 0.5847 0.3388 0.5266 0.5265 0.8951 0.9003 11.1929 16 6 16.4643 28.5714
0.4041 44.0 2728 2.0105 0.5844 0.3468 0.5257 0.5256 0.8957 0.901 11.1786 15 6 16.5357 29.2857
0.3925 45.0 2790 2.0220 0.5787 0.3423 0.5179 0.5185 0.8936 0.8992 11.25 16 6 16.5143 32.1429
0.4095 46.0 2852 2.0179 0.581 0.3404 0.5197 0.5202 0.8957 0.8998 11.2143 16 6 16.4357 29.2857
0.397 47.0 2914 2.0124 0.5803 0.3385 0.5188 0.5193 0.8952 0.899 11.2357 16 6 16.2786 32.1429
0.3801 48.0 2976 2.0186 0.5778 0.3359 0.518 0.518 0.8944 0.8986 11.2143 16 6 16.4 32.1429
0.3966 49.0 3038 2.0234 0.5807 0.337 0.5185 0.5192 0.8953 0.9001 11.2571 16 6 16.4929 30.0
0.3838 50.0 3100 2.0317 0.5807 0.3427 0.523 0.5234 0.8954 0.8989 11.0571 16 6 16.2786 26.4286
0.3818 51.0 3162 2.0281 0.5811 0.3428 0.5238 0.5242 0.8956 0.9001 11.1643 16 6 16.3643 30.7143
0.3793 52.0 3224 2.0399 0.5824 0.3438 0.5214 0.522 0.8947 0.9003 11.2071 16 6 16.4714 30.7143
0.3734 53.0 3286 2.0470 0.5811 0.3413 0.5222 0.5227 0.8952 0.9 11.1643 15 6 16.4214 29.2857
0.3876 54.0 3348 2.0509 0.5764 0.3382 0.515 0.5156 0.8948 0.8983 11.1071 15 6 16.2643 28.5714
0.3736 55.0 3410 2.0498 0.5722 0.3331 0.5135 0.514 0.8937 0.8972 11.1357 16 6 16.3071 27.8571
0.3981 56.0 3472 2.0499 0.5726 0.3337 0.5133 0.5138 0.8939 0.8977 11.1286 15 6 16.3357 29.2857
0.3731 57.0 3534 2.0500 0.5767 0.3353 0.5173 0.5176 0.8946 0.8984 11.1286 15 6 16.3643 27.8571
0.3786 58.0 3596 2.0529 0.5779 0.3377 0.5199 0.5208 0.895 0.8994 11.1929 16 6 16.4357 28.5714
0.3648 59.0 3658 2.0545 0.5766 0.3348 0.518 0.5181 0.8939 0.8985 11.2143 15 6 16.5143 30.0
0.373 60.0 3720 2.0548 0.5772 0.3353 0.5189 0.5189 0.8941 0.8987 11.2143 15 6 16.5071 30.0

Framework versions

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