<!-- 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. -->
stance_detection
This model is a fine-tuned version of bert-base-cased towards 26 US SPAC stock mergers on Twitter. It achieves the following results on the evaluation set:
- Loss: 0.4906
- Accuracy: 0.8409
- F1w: 0.8574
- Acc0: 0.8293
- Acc1: 0.6
- Acc2: 0.7652
- Acc3: 0.8637
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: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1w | Acc0 | Acc1 | Acc2 | Acc3 |
|---|---|---|---|---|---|---|---|---|---|
| 0.7748 | 1.0 | 194 | 0.5172 | 0.8158 | 0.8297 | 0.8699 | 0.0 | 0.7429 | 0.8248 |
| 0.5181 | 2.0 | 388 | 0.4692 | 0.8509 | 0.8587 | 0.8699 | 0.4 | 0.7429 | 0.8743 |
| 0.3868 | 3.0 | 582 | 0.4906 | 0.8409 | 0.8574 | 0.8293 | 0.6 | 0.7652 | 0.8637 |
Framework versions
- Transformers 4.22.1
- Pytorch 1.11.0
- Datasets 2.5.2
- Tokenizers 0.12.1
