tiny-mlm-glue-qnli-target-glue-qqp
This model is a fine-tuned version of muhtasham/tiny-mlm-glue-qnli on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4125
- Accuracy: 0.7971
- F1: 0.7707
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: 3e-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: constant
- num_epochs: 200
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
---|---|---|---|---|---|
0.5776 | 0.04 | 500 | 0.5177 | 0.7272 | 0.6831 |
0.5088 | 0.09 | 1000 | 0.4828 | 0.7515 | 0.7055 |
0.4952 | 0.13 | 1500 | 0.4939 | 0.7383 | 0.7143 |
0.4797 | 0.18 | 2000 | 0.4681 | 0.7547 | 0.7225 |
0.4723 | 0.22 | 2500 | 0.4564 | 0.7621 | 0.7274 |
0.4551 | 0.26 | 3000 | 0.4475 | 0.7693 | 0.7351 |
0.4573 | 0.31 | 3500 | 0.4479 | 0.7676 | 0.7372 |
0.4496 | 0.35 | 4000 | 0.4483 | 0.7668 | 0.7390 |
0.4503 | 0.4 | 4500 | 0.4413 | 0.7720 | 0.7436 |
0.4407 | 0.44 | 5000 | 0.4192 | 0.7899 | 0.7498 |
0.4288 | 0.48 | 5500 | 0.4261 | 0.7845 | 0.7512 |
0.4292 | 0.53 | 6000 | 0.4058 | 0.8022 | 0.7581 |
0.4235 | 0.57 | 6500 | 0.4201 | 0.7893 | 0.7560 |
0.4251 | 0.62 | 7000 | 0.4050 | 0.8007 | 0.7593 |
0.4161 | 0.66 | 7500 | 0.4063 | 0.8040 | 0.7652 |
0.4297 | 0.7 | 8000 | 0.4116 | 0.7959 | 0.7617 |
0.4201 | 0.75 | 8500 | 0.3975 | 0.8069 | 0.7677 |
0.4142 | 0.79 | 9000 | 0.4186 | 0.7889 | 0.7609 |
0.4113 | 0.84 | 9500 | 0.3900 | 0.8112 | 0.7687 |
0.413 | 0.88 | 10000 | 0.3852 | 0.8161 | 0.7732 |
0.4084 | 0.92 | 10500 | 0.3826 | 0.8161 | 0.7714 |
0.4083 | 0.97 | 11000 | 0.3826 | 0.8187 | 0.7733 |
0.4057 | 1.01 | 11500 | 0.4016 | 0.8029 | 0.7711 |
0.3846 | 1.06 | 12000 | 0.3803 | 0.8187 | 0.7759 |
0.3949 | 1.1 | 12500 | 0.3827 | 0.8154 | 0.7773 |
0.3823 | 1.14 | 13000 | 0.3878 | 0.8136 | 0.7763 |
0.3717 | 1.19 | 13500 | 0.4125 | 0.7971 | 0.7707 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2
- Downloads last month
- 8