metadata
tags:
- paraphrasing
- generated_from_trainer
datasets:
- paws
metrics:
- rouge
model-index:
- name: pegasus-xsum-finetuned-paws
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: paws
type: paws
args: labeled_final
metrics:
- name: Rouge1
type: rouge
value: 92.4371
pegasus-xsum-finetuned-paws
This model is a fine-tuned version of google/pegasus-xsum on the paws dataset. It achieves the following results on the evaluation set:
- Loss: 2.1199
- Rouge1: 92.4371
- Rouge2: 75.4061
- Rougel: 84.1519
- Rougelsum: 84.1958
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: 32
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
- mixed_precision_training: Native AMP
- label_smoothing_factor: 0.1
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|---|---|---|---|---|---|---|---|
| 2.1481 | 1.46 | 1000 | 2.0112 | 93.7727 | 73.3021 | 84.2963 | 84.2506 |
| 2.0113 | 2.93 | 2000 | 2.0579 | 93.813 | 73.4119 | 84.3674 | 84.2693 |
| 2.054 | 4.39 | 3000 | 2.0890 | 93.3926 | 73.3727 | 84.2814 | 84.1649 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.11.0
- Datasets 2.1.0
- Tokenizers 0.12.1