Hugging Face's logo

arthurmluz
/
ptt5-xlsumm-gptextsum

Text2Text Generation
Transformers
PyTorch
t5
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Community
ptt5-xlsumm-gptextsum / README.md
arthurmluz's picture
arthurmluz
Model save
e4eca9c
preview code
|
raw
history blame
4.48 kB
metadata
license: mit
base_model: arthurmluz/ptt5-xlsumm-30epochs
tags:
  - generated_from_trainer
metrics:
  - rouge
model-index:
  - name: ptt5-xlsumm-gptextsum
    results: []

ptt5-xlsumm-gptextsum

This model is a fine-tuned version of arthurmluz/ptt5-xlsumm-30epochs on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 2.2170
  • Rouge1: 0.4041
  • Rouge2: 0.207
  • Rougel: 0.3276
  • Rougelsum: 0.3632
  • Gen Len: 18.875

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: 2e-05
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Rouge1 Rouge2 Rougel Rougelsum Gen Len
No log 1.0 70 2.3635 0.318 0.1516 0.2478 0.2858 18.675
No log 2.0 140 2.1986 0.3333 0.1619 0.268 0.3003 18.7
2.3332 3.0 210 2.1209 0.3244 0.1429 0.2551 0.293 18.5
2.3332 4.0 280 2.0820 0.3448 0.1618 0.2703 0.3098 18.5
2.3332 5.0 350 2.0565 0.3502 0.1769 0.2756 0.3135 18.75
1.785 6.0 420 2.0446 0.3541 0.1808 0.2734 0.3169 18.825
1.785 7.0 490 2.0435 0.3628 0.1867 0.2843 0.3228 18.75
1.785 8.0 560 2.0482 0.3679 0.1886 0.2918 0.3314 18.725
1.4946 9.0 630 2.0481 0.3741 0.1869 0.2949 0.3336 18.85
1.4946 10.0 700 2.0494 0.379 0.1979 0.3012 0.338 18.8
1.4946 11.0 770 2.0602 0.3786 0.196 0.3045 0.3342 18.8
1.2836 12.0 840 2.0691 0.3749 0.1971 0.3012 0.3352 18.85
1.2836 13.0 910 2.0795 0.3789 0.1953 0.3082 0.3371 18.8
1.2836 14.0 980 2.0898 0.3809 0.196 0.3052 0.3378 18.825
1.1472 15.0 1050 2.1076 0.3904 0.1979 0.3137 0.3462 18.875
1.1472 16.0 1120 2.1109 0.3905 0.1929 0.3101 0.3436 18.875
1.1472 17.0 1190 2.1253 0.4046 0.2029 0.3242 0.3594 18.875
1.0175 18.0 1260 2.1418 0.4093 0.2074 0.3255 0.3614 18.875
1.0175 19.0 1330 2.1578 0.4065 0.213 0.3289 0.363 18.875
0.9398 20.0 1400 2.1593 0.4125 0.2197 0.333 0.3716 18.875
0.9398 21.0 1470 2.1663 0.4091 0.2111 0.3282 0.368 18.875
0.9398 22.0 1540 2.1775 0.4081 0.2111 0.3301 0.366 18.875
0.8644 23.0 1610 2.1889 0.4015 0.206 0.3274 0.3613 18.875
0.8644 24.0 1680 2.1932 0.4015 0.206 0.3274 0.3613 18.875
0.8644 25.0 1750 2.2016 0.4075 0.2089 0.331 0.3672 18.875
0.8385 26.0 1820 2.2058 0.4082 0.2087 0.33 0.3688 18.875
0.8385 27.0 1890 2.2089 0.4075 0.2089 0.331 0.3672 18.875
0.8385 28.0 1960 2.2131 0.4061 0.2079 0.3286 0.3651 18.875
0.7999 29.0 2030 2.2161 0.4061 0.2079 0.3286 0.3651 18.875
0.7999 30.0 2100 2.2170 0.4041 0.207 0.3276 0.3632 18.875

Framework versions

  • Transformers 4.34.0
  • Pytorch 2.0.1+cu117
  • Datasets 2.14.5
  • Tokenizers 0.14.1