Edit model card

falcon-summ

This model is a fine-tuned version of Falconsai/text_summarization on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 2.1784
  • Rouge1: 0.1921
  • Rouge2: 0.0958
  • Rougel: 0.1642
  • Rougelsum: 0.1643
  • Gen Len: 19.0

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: 2e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 30
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Rouge1 Rouge2 Rougel Rougelsum Gen Len
No log 1.0 62 2.6597 0.1336 0.0451 0.1109 0.1109 19.0
No log 2.0 124 2.5126 0.1519 0.0584 0.1258 0.1259 19.0
No log 3.0 186 2.4354 0.1702 0.0705 0.142 0.1422 19.0
No log 4.0 248 2.3855 0.1915 0.0922 0.1616 0.1618 19.0
No log 5.0 310 2.3499 0.1932 0.0939 0.1641 0.1644 19.0
No log 6.0 372 2.3251 0.1951 0.0959 0.1665 0.1667 19.0
No log 7.0 434 2.3035 0.1943 0.0963 0.1655 0.1656 19.0
No log 8.0 496 2.2879 0.1941 0.0953 0.1652 0.1653 19.0
2.6368 9.0 558 2.2695 0.1952 0.0958 0.1665 0.1667 19.0
2.6368 10.0 620 2.2588 0.194 0.0952 0.1651 0.1653 19.0
2.6368 11.0 682 2.2476 0.1954 0.0968 0.1668 0.167 19.0
2.6368 12.0 744 2.2396 0.1944 0.0965 0.1656 0.1657 19.0
2.6368 13.0 806 2.2296 0.1929 0.0963 0.1641 0.1644 19.0
2.6368 14.0 868 2.2242 0.1928 0.0969 0.1639 0.1639 19.0
2.6368 15.0 930 2.2159 0.1935 0.0971 0.164 0.1641 19.0
2.6368 16.0 992 2.2114 0.1924 0.0965 0.164 0.1641 19.0
2.3506 17.0 1054 2.2077 0.1926 0.0973 0.1644 0.1645 19.0
2.3506 18.0 1116 2.2031 0.1933 0.0968 0.1647 0.1648 19.0
2.3506 19.0 1178 2.1971 0.1928 0.0962 0.1643 0.1644 19.0
2.3506 20.0 1240 2.1956 0.1925 0.0956 0.165 0.1651 19.0
2.3506 21.0 1302 2.1903 0.1927 0.0958 0.1644 0.1644 19.0
2.3506 22.0 1364 2.1882 0.1933 0.0972 0.1653 0.1653 19.0
2.3506 23.0 1426 2.1858 0.1921 0.0956 0.1639 0.1641 19.0
2.3506 24.0 1488 2.1842 0.1921 0.0956 0.1642 0.1643 19.0
2.2758 25.0 1550 2.1832 0.1919 0.0958 0.1645 0.1647 19.0
2.2758 26.0 1612 2.1815 0.1922 0.0958 0.1646 0.1647 19.0
2.2758 27.0 1674 2.1795 0.1924 0.0962 0.1646 0.1647 19.0
2.2758 28.0 1736 2.1790 0.1922 0.0961 0.1646 0.1647 19.0
2.2758 29.0 1798 2.1784 0.1925 0.0963 0.1645 0.1646 19.0
2.2758 30.0 1860 2.1784 0.1921 0.0958 0.1642 0.1643 19.0

Framework versions

  • Transformers 4.39.3
  • Pytorch 2.1.2
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
1
Safetensors
Model size
60.5M params
Tensor type
F32
·
Inference API
This model can be loaded on Inference API (serverless).

Finetuned from