<!-- 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. -->
scenario-no_kd_weight_reset-data-smsa-model-xlm-roberta-base
This model is a fine-tuned version of xlm-roberta-base on the smsa dataset. It achieves the following results on the evaluation set:
- Loss: 0.5261
- Accuracy: 0.8786
- F1: 0.8368
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: 6969
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
---|---|---|---|---|---|
No log | 0.29 | 100 | 0.6517 | 0.7254 | 0.4930 |
No log | 0.58 | 200 | 0.4771 | 0.8238 | 0.7219 |
No log | 0.87 | 300 | 0.4506 | 0.8246 | 0.7348 |
No log | 1.16 | 400 | 0.4853 | 0.8294 | 0.7669 |
0.526 | 1.45 | 500 | 0.3841 | 0.8444 | 0.7686 |
0.526 | 1.74 | 600 | 0.3501 | 0.8746 | 0.8393 |
0.526 | 2.03 | 700 | 0.3767 | 0.8722 | 0.8293 |
0.526 | 2.33 | 800 | 0.3945 | 0.8619 | 0.8198 |
0.526 | 2.62 | 900 | 0.3487 | 0.8738 | 0.8380 |
0.2817 | 2.91 | 1000 | 0.3619 | 0.8722 | 0.8338 |
0.2817 | 3.2 | 1100 | 0.4189 | 0.8651 | 0.8360 |
0.2817 | 3.49 | 1200 | 0.4260 | 0.8603 | 0.8170 |
0.2817 | 3.78 | 1300 | 0.3946 | 0.8619 | 0.7986 |
0.2817 | 4.07 | 1400 | 0.4748 | 0.8619 | 0.8294 |
0.1958 | 4.36 | 1500 | 0.5528 | 0.8492 | 0.8106 |
0.1958 | 4.65 | 1600 | 0.5588 | 0.8563 | 0.8007 |
0.1958 | 4.94 | 1700 | 0.7266 | 0.8357 | 0.7780 |
0.1958 | 5.23 | 1800 | 0.8070 | 0.8405 | 0.7969 |
0.1958 | 5.52 | 1900 | 0.5910 | 0.8643 | 0.8225 |
0.1457 | 5.81 | 2000 | 0.6039 | 0.8730 | 0.8253 |
0.1457 | 6.1 | 2100 | 0.5261 | 0.8786 | 0.8368 |
Framework versions
- Transformers 4.33.3
- Pytorch 2.0.1
- Datasets 2.14.5
- Tokenizers 0.13.3