metadata
library_name: transformers
license: mit
base_model: openai-community/gpt2-medium
tags:
- generated_from_trainer
model-index:
- name: tinystories_scrambled_words
results: []
tinystories_scrambled_words
This model is a fine-tuned version of openai-community/gpt2-medium on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 3.7107
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: 0.0001
- 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: linear
- lr_scheduler_warmup_steps: 1
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.3727 | 0.1051 | 400 | 4.3237 |
| 4.1849 | 0.2102 | 800 | 4.1539 |
| 4.1369 | 0.3153 | 1200 | 4.0841 |
| 4.0574 | 0.4204 | 1600 | 4.0165 |
| 4.0123 | 0.5255 | 2000 | 3.9716 |
| 3.9874 | 0.6306 | 2400 | 3.9359 |
| 3.9459 | 0.7357 | 2800 | 3.9111 |
| 3.9004 | 0.8408 | 3200 | 3.8865 |
| 3.9493 | 0.9459 | 3600 | 3.8587 |
| 3.8599 | 1.0510 | 4000 | 3.8427 |
| 3.8753 | 1.1561 | 4400 | 3.8281 |
| 3.8352 | 1.2612 | 4800 | 3.8156 |
| 3.8194 | 1.3663 | 5200 | 3.8007 |
| 3.8334 | 1.4714 | 5600 | 3.7884 |
| 3.8279 | 1.5765 | 6000 | 3.7814 |
| 3.8233 | 1.6816 | 6400 | 3.7728 |
| 3.7066 | 1.7867 | 6800 | 3.7629 |
| 3.7625 | 1.8918 | 7200 | 3.7544 |
| 3.7696 | 1.9969 | 7600 | 3.7465 |
| 3.7041 | 2.1020 | 8000 | 3.7424 |
| 3.7223 | 2.2071 | 8400 | 3.7369 |
| 3.7445 | 2.3122 | 8800 | 3.7362 |
| 3.7023 | 2.4173 | 9200 | 3.7291 |
| 3.6926 | 2.5224 | 9600 | 3.7284 |
| 3.7196 | 2.6275 | 10000 | 3.7212 |
| 3.7469 | 2.7326 | 10400 | 3.7183 |
| 3.7082 | 2.8377 | 10800 | 3.7159 |
| 3.7284 | 2.9428 | 11200 | 3.7147 |
Framework versions
- Transformers 4.44.1
- Pytorch 2.2.2
- Datasets 2.18.0
- Tokenizers 0.19.1