<!-- 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. -->
gpt-expt-sp-v3-K-600-kmeans-v6
This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0166
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
- gradient_accumulation_steps: 8
- total_train_batch_size: 512
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.1506 | 17.92 | 5000 | 0.0679 |
| 0.0425 | 35.84 | 10000 | 0.0358 |
| 0.0236 | 53.76 | 15000 | 0.0204 |
| 0.0461 | 71.68 | 20000 | 0.1917 |
| 0.027 | 89.61 | 25000 | 0.0181 |
| 0.0187 | 107.53 | 30000 | 0.0177 |
| 0.0208 | 125.45 | 35000 | 0.0171 |
| 0.0184 | 143.37 | 40000 | 0.0170 |
| 0.0177 | 161.29 | 45000 | 0.0171 |
| 0.0182 | 179.21 | 50000 | 0.0170 |
| 0.0175 | 197.13 | 55000 | 0.0169 |
| 0.0174 | 215.05 | 60000 | 0.0169 |
| 0.0173 | 232.97 | 65000 | 0.0169 |
| 0.0172 | 250.9 | 70000 | 0.0168 |
| 0.0171 | 268.82 | 75000 | 0.0168 |
| 0.017 | 286.74 | 80000 | 0.0167 |
| 0.0169 | 304.66 | 85000 | 0.0167 |
| 0.0169 | 322.58 | 90000 | 0.0167 |
| 0.0168 | 340.5 | 95000 | 0.0167 |
| 0.0168 | 358.42 | 100000 | 0.0166 |
| 0.0167 | 376.34 | 105000 | 0.0166 |
| 0.0167 | 394.27 | 110000 | 0.0166 |
| 0.0167 | 412.19 | 115000 | 0.0166 |
| 0.0166 | 430.11 | 120000 | 0.0166 |
| 0.0166 | 448.03 | 125000 | 0.0166 |
| 0.0166 | 465.95 | 130000 | 0.0166 |
| 0.0166 | 483.87 | 135000 | 0.0166 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
