metadata
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- accuracy
widget:
- 全国高校科研质量排名,清华北大浙大无缘前五。
model-index:
- name: classification_tnews_scarce_imbalanced
results: []
classification_tnews_scarce_imbalanced
This model is a fine-tuned version of hfl/chinese-roberta-wwm-ext on a very small & imbalanced subset of TNews dataset. It achieves the following results on the evaluation set:
- Loss: 1.5812
- Accuracy: 0.5867
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: 10
- eval_batch_size: 10
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
2.4743 | 1.0 | 15 | 2.5030 | 0.24 |
2.1581 | 2.0 | 30 | 2.2729 | 0.4533 |
1.7781 | 3.0 | 45 | 2.0809 | 0.5267 |
1.5336 | 4.0 | 60 | 1.9246 | 0.5333 |
1.2713 | 5.0 | 75 | 1.8133 | 0.54 |
1.1 | 6.0 | 90 | 1.7276 | 0.5667 |
0.9569 | 7.0 | 105 | 1.6676 | 0.5733 |
0.8755 | 8.0 | 120 | 1.6177 | 0.5867 |
0.7461 | 9.0 | 135 | 1.5917 | 0.5867 |
0.7229 | 10.0 | 150 | 1.5812 | 0.5867 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.0+cu116
- Datasets 2.8.0
- Tokenizers 0.13.2