<!-- 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. -->
gpt2_left_out_qed
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 3.9486
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: 0.0005
- 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: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
5.9695 | 0.27 | 500 | 5.0679 |
4.7417 | 0.53 | 1000 | 4.6811 |
4.4136 | 0.8 | 1500 | 4.4369 |
4.2076 | 1.06 | 2000 | 4.2985 |
4.0279 | 1.33 | 2500 | 4.2048 |
3.9505 | 1.59 | 3000 | 4.1137 |
3.8781 | 1.86 | 3500 | 4.0482 |
3.7338 | 2.12 | 4000 | 4.0046 |
3.6392 | 2.39 | 4500 | 3.9628 |
3.6228 | 2.65 | 5000 | 3.9115 |
3.5944 | 2.92 | 5500 | 3.8738 |
3.4222 | 3.18 | 6000 | 3.8797 |
3.3836 | 3.45 | 6500 | 3.8576 |
3.3995 | 3.71 | 7000 | 3.8251 |
3.3827 | 3.98 | 7500 | 3.7995 |
3.1568 | 4.24 | 8000 | 3.8348 |
3.1778 | 4.51 | 8500 | 3.8171 |
3.1853 | 4.77 | 9000 | 3.7963 |
3.1451 | 5.04 | 9500 | 3.8059 |
2.9278 | 5.31 | 10000 | 3.8298 |
2.9608 | 5.57 | 10500 | 3.8176 |
2.9762 | 5.84 | 11000 | 3.8047 |
2.8716 | 6.1 | 11500 | 3.8433 |
2.7239 | 6.37 | 12000 | 3.8523 |
2.7435 | 6.63 | 12500 | 3.8541 |
2.7524 | 6.9 | 13000 | 3.8446 |
2.6032 | 7.16 | 13500 | 3.8854 |
2.5322 | 7.43 | 14000 | 3.8967 |
2.5369 | 7.69 | 14500 | 3.8983 |
2.5467 | 7.96 | 15000 | 3.8966 |
2.3979 | 8.22 | 15500 | 3.9284 |
2.3767 | 8.49 | 16000 | 3.9334 |
2.3852 | 8.75 | 16500 | 3.9357 |
2.3805 | 9.02 | 17000 | 3.9395 |
2.3012 | 9.28 | 17500 | 3.9463 |
2.3044 | 9.55 | 18000 | 3.9484 |
2.3007 | 9.81 | 18500 | 3.9486 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3