LED_sum_challenge

This model is a fine-tuned version of allenai/led-base-16384 on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 3.8042
  • Rouge1: 0.2495
  • Rouge2: 0.0724
  • Rougel: 0.1912
  • Rougelsum: 0.192
  • Gen Len: 20.5
  • Bleu: 0.0232
  • Precisions: 0.0926
  • Brevity Penalty: 0.6016
  • Length Ratio: 0.6631
  • Translation Length: 801.0
  • Reference Length: 1208.0
  • Precision: 0.8797
  • Recall: 0.8676
  • F1: 0.8736
  • Hashcode: roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)

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: 4
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 16
  • optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Rouge1 Rouge2 Rougel Rougelsum Gen Len Bleu Precisions Brevity Penalty Length Ratio Translation Length Reference Length Precision Recall F1 Hashcode
No log 1.0 7 8.1739 0.2255 0.0527 0.1686 0.1688 21.0 0.0157 0.069 0.6607 0.707 854.0 1208.0 0.8668 0.8574 0.862 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)
No log 2.0 14 6.9457 0.2251 0.0588 0.1702 0.1685 20.7 0.0171 0.0737 0.6408 0.6921 836.0 1208.0 0.8737 0.8597 0.8666 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)
No log 3.0 21 5.4862 0.2391 0.0632 0.181 0.1805 20.52 0.021 0.0825 0.6431 0.6937 838.0 1208.0 0.8798 0.862 0.8708 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)
No log 4.0 28 4.7435 0.243 0.0758 0.1901 0.1892 20.72 0.0266 0.0886 0.6095 0.6689 808.0 1208.0 0.8775 0.8662 0.8717 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)
No log 5.0 35 4.3805 0.2557 0.0788 0.1924 0.1921 20.48 0.0248 0.1003 0.5857 0.6515 787.0 1208.0 0.8811 0.8686 0.8747 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)
No log 6.0 42 4.1441 0.2485 0.0701 0.1886 0.1894 20.52 0.0209 0.0929 0.5982 0.6606 798.0 1208.0 0.8816 0.868 0.8747 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)
No log 7.0 49 3.9952 0.2574 0.0713 0.1994 0.1997 20.54 0.0213 0.0954 0.6073 0.6672 806.0 1208.0 0.8811 0.8689 0.8749 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)
No log 8.0 56 3.8994 0.2524 0.067 0.192 0.192 20.58 0.0203 0.0908 0.614 0.6722 812.0 1208.0 0.8782 0.8675 0.8727 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)
No log 9.0 63 3.8355 0.2512 0.0676 0.1917 0.1925 20.54 0.0201 0.0901 0.6062 0.6664 805.0 1208.0 0.8793 0.8681 0.8736 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)
No log 10.0 70 3.8042 0.2495 0.0724 0.1912 0.192 20.5 0.0232 0.0926 0.6016 0.6631 801.0 1208.0 0.8797 0.8676 0.8736 roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.53.0)

Framework versions

  • Transformers 4.53.0
  • Pytorch 2.7.0+cu126
  • Datasets 3.6.0
  • Tokenizers 0.21.1
Downloads last month
4
Safetensors
Model size
162M params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for floflodebilbao/LED_sum_challenge

Finetuned
(37)
this model