---
tags:
- generated_from_trainer
datasets:
- null
metrics:
- rouge
model_index:
- name: bart-base-finetuned-xsum
  results:
  - task:
      name: Sequence-to-sequence Language Modeling
      type: text2text-generation
    metric:
      name: Rouge1
      type: rouge
      value: 27.8379
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# bart-base-finetuned-xsum

This model is a fine-tuned version of [facebook/bart-base](https://huggingface.co/facebook/bart-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5671
- Rouge1: 27.8379
- Rouge2: 16.2683
- Rougel: 24.1898
- Rougelsum: 25.5234
- Gen Len: 20.0

## 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: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1

### Training results

| Training Loss | Epoch | Step | Validation Loss | Rouge1  | Rouge2  | Rougel  | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| 1.9931        | 1.0   | 879  | 1.5671          | 27.8379 | 16.2683 | 24.1898 | 25.5234   | 20.0    |


### Framework versions

- Transformers 4.9.2
- Pytorch 1.9.0+cu102
- Datasets 1.11.0
- Tokenizers 0.10.3