|
--- |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- rouge |
|
model-index: |
|
- name: distill-pegasus-cnn-arxiv-pubmed-v3-e8 |
|
results: [] |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# distill-pegasus-cnn-arxiv-pubmed-v3-e8 |
|
|
|
This model is a fine-tuned version of [theojolliffe/distill-pegasus-cnn-arxiv-pubmed](https://huggingface.co/theojolliffe/distill-pegasus-cnn-arxiv-pubmed) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.6844 |
|
- Rouge1: 49.0081 |
|
- Rouge2: 30.6784 |
|
- Rougel: 33.5258 |
|
- Rougelsum: 45.5354 |
|
- Gen Len: 125.6852 |
|
|
|
## 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: 1 |
|
- eval_batch_size: 1 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 8 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len | |
|
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:--------:| |
|
| 2.7633 | 1.0 | 795 | 2.1211 | 48.9615 | 30.3509 | 33.7359 | 44.508 | 124.7963 | |
|
| 2.3051 | 2.0 | 1590 | 1.9464 | 48.6806 | 30.452 | 34.2187 | 44.6379 | 124.6296 | |
|
| 2.2244 | 3.0 | 2385 | 1.8294 | 48.9739 | 30.6717 | 33.605 | 45.0942 | 125.3704 | |
|
| 2.0733 | 4.0 | 3180 | 1.7769 | 49.0049 | 30.8354 | 33.6965 | 44.8603 | 125.7037 | |
|
| 1.9759 | 5.0 | 3975 | 1.7192 | 50.3946 | 32.1072 | 34.5453 | 46.4493 | 125.5741 | |
|
| 1.9478 | 6.0 | 4770 | 1.7037 | 49.4631 | 31.654 | 34.4601 | 46.2376 | 125.5185 | |
|
| 1.9016 | 7.0 | 5565 | 1.6874 | 48.2641 | 29.6354 | 33.1059 | 44.8436 | 125.6852 | |
|
| 1.8882 | 8.0 | 6360 | 1.6844 | 49.0081 | 30.6784 | 33.5258 | 45.5354 | 125.6852 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.18.0 |
|
- Pytorch 1.11.0+cu113 |
|
- Datasets 2.1.0 |
|
- Tokenizers 0.12.1 |
|
|
