roberta-base-mnli_MULTI
This model is a fine-tuned version of WillHeld/roberta-base-mnli on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.6448
- Acc: 0.8397
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: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
Training results
Training Loss | Epoch | Step | Validation Loss | Acc |
---|---|---|---|---|
0.4594 | 0.33 | 2000 | 0.4535 | 0.8292 |
0.4231 | 0.67 | 4000 | 0.4440 | 0.8323 |
0.4128 | 1.0 | 6000 | 0.4484 | 0.8352 |
0.3137 | 1.33 | 8000 | 0.4905 | 0.8338 |
0.3168 | 1.66 | 10000 | 0.4740 | 0.8372 |
0.3176 | 2.0 | 12000 | 0.4624 | 0.8415 |
0.2347 | 2.33 | 14000 | 0.5300 | 0.8361 |
0.2361 | 2.66 | 16000 | 0.5097 | 0.8378 |
0.238 | 3.0 | 18000 | 0.5016 | 0.8407 |
0.18 | 3.33 | 20000 | 0.5873 | 0.8382 |
0.179 | 3.66 | 22000 | 0.5917 | 0.8374 |
0.1786 | 4.0 | 24000 | 0.5868 | 0.8398 |
0.1409 | 4.33 | 26000 | 0.6468 | 0.8371 |
0.1395 | 4.66 | 28000 | 0.6423 | 0.8371 |
0.1401 | 4.99 | 30000 | 0.6448 | 0.8397 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.13.0+cu117
- Datasets 2.7.1
- Tokenizers 0.12.1
- Downloads last month
- 3