File size: 1,679 Bytes
1e7bf4c 1b32bb4 1e7bf4c 1b32bb4 1e7bf4c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 |
---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: distilroberta-clickbait
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. -->
# distilroberta-clickbait
This model is a fine-tuned version of [distilroberta-base](https://huggingface.co/distilroberta-base) on a dataset of headlines.
It achieves the following results on the evaluation set:
- Loss: 0.0268
- Acc: 0.9963
## Training and evaluation data
The following data sources were used:
* 32k headlines classified as clickbait/not-clickbait from [kaggle](https://www.kaggle.com/amananandrai/clickbait-dataset)
* A dataset of headlines from https://github.com/MotiBaadror/Clickbait-Detection
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 12345
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 16
- num_epochs: 20
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Acc |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.0195 | 1.0 | 981 | 0.0192 | 0.9954 |
| 0.0026 | 2.0 | 1962 | 0.0172 | 0.9963 |
| 0.0031 | 3.0 | 2943 | 0.0275 | 0.9945 |
| 0.0003 | 4.0 | 3924 | 0.0268 | 0.9963 |
### Framework versions
- Transformers 4.11.3
- Pytorch 1.10.1
- Datasets 1.17.0
- Tokenizers 0.10.3
|