metadata
license: mit
base_model: gpt2
tags:
- generated_from_trainer
model-index:
- name: vi_gpt
results: []
vi_gpt
This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 3.0359
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
- 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: 100
- training_steps: 2561
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.7072 | 0.0200 | 256 | 4.4571 |
| 4.0563 | 0.0400 | 512 | 3.9044 |
| 3.7398 | 0.0600 | 768 | 3.6152 |
| 3.569 | 0.0800 | 1024 | 3.4209 |
| 3.4202 | 0.1001 | 1280 | 3.2897 |
| 3.319 | 0.1201 | 1536 | 3.1876 |
| 3.2328 | 0.1401 | 1792 | 3.1171 |
| 3.1905 | 0.1601 | 2048 | 3.0668 |
| 3.1555 | 0.1801 | 2304 | 3.0411 |
| 3.1468 | 0.2001 | 2560 | 3.0359 |
Framework versions
- Transformers 4.42.3
- Pytorch 2.2.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1