<!-- 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. -->
codeparrot-ds-sample-2ep-batchsize32
This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.5721
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: 32
- eval_batch_size: 32
- seed: 42
- distributed_type: tpu
- gradient_accumulation_steps: 8
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 2
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.3529 | 0.19 | 1000 | 2.8073 |
| 2.4602 | 0.37 | 2000 | 2.2907 |
| 2.1127 | 0.56 | 3000 | 2.0745 |
| 1.9187 | 0.74 | 4000 | 1.9287 |
| 1.782 | 0.93 | 5000 | 1.8234 |
| 1.639 | 1.11 | 6000 | 1.7456 |
| 1.5519 | 1.3 | 7000 | 1.6738 |
| 1.489 | 1.49 | 8000 | 1.6235 |
| 1.4372 | 1.67 | 9000 | 1.5874 |
| 1.4122 | 1.86 | 10000 | 1.5721 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.11.0+cu102
- Datasets 2.1.0
- Tokenizers 0.12.1
