Edit model card

bart-abs-2409-1947-lr-0.0003-bs-2-maxep-10

This model is a fine-tuned version of sshleifer/distilbart-xsum-12-6 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 7.4055
  • Rouge/rouge1: 0.2434
  • Rouge/rouge2: 0.0406
  • Rouge/rougel: 0.2027
  • Rouge/rougelsum: 0.203
  • Bertscore/bertscore-precision: 0.8533
  • Bertscore/bertscore-recall: 0.8523
  • Bertscore/bertscore-f1: 0.8528
  • Meteor: 0.194
  • Gen Len: 34.0

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: 0.0003
  • 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: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Rouge/rouge1 Rouge/rouge2 Rouge/rougel Rouge/rougelsum Bertscore/bertscore-precision Bertscore/bertscore-recall Bertscore/bertscore-f1 Meteor Gen Len
0.6036 1.0 434 6.4324 0.2828 0.0721 0.2287 0.229 0.8776 0.863 0.8702 0.1865 29.0
0.6828 2.0 868 5.8330 0.2817 0.0594 0.2119 0.2118 0.8553 0.8592 0.8572 0.2216 38.0
0.5861 3.0 1302 6.4604 0.2456 0.0556 0.1909 0.1914 0.8622 0.8598 0.861 0.2363 41.0
0.4888 4.0 1736 6.3098 0.2591 0.0557 0.2009 0.2012 0.859 0.8605 0.8597 0.2511 47.0
0.4314 5.0 2170 6.7076 0.3042 0.0843 0.2354 0.2357 0.8695 0.8605 0.8649 0.22 32.0
0.379 6.0 2604 6.7173 0.2844 0.0777 0.2203 0.22 0.8641 0.8631 0.8635 0.2167 34.0
0.3245 7.0 3038 6.8382 0.2613 0.0803 0.214 0.2142 0.8711 0.8469 0.8588 0.2102 25.0
0.289 8.0 3472 7.0253 0.2396 0.0481 0.1978 0.198 0.8526 0.8544 0.8534 0.2079 34.0
0.2575 9.0 3906 7.1264 0.3089 0.0693 0.2341 0.2343 0.8686 0.8645 0.8665 0.2205 36.0
0.2312 10.0 4340 7.4055 0.2434 0.0406 0.2027 0.203 0.8533 0.8523 0.8528 0.194 34.0

Framework versions

  • Transformers 4.44.2
  • Pytorch 2.4.0
  • Datasets 3.0.0
  • Tokenizers 0.19.1
Downloads last month
6
Safetensors
Model size
306M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for roequitz/bart-abs-2409-1947-lr-0.0003-bs-2-maxep-10

Finetuned
(43)
this model