metadata
base_model: ai-forever/rugpt3medium_based_on_gpt2
tags:
- generated_from_trainer
model-index:
- name: my_rugpt3medium_finetune
results: []
my_rugpt3medium_finetune
This model is a fine-tuned version of ai-forever/rugpt3medium_based_on_gpt2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.9269
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: 1e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 24
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 25
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 3.601 | 1.6 | 25 | 3.6157 |
| 3.601 | 3.19 | 50 | 3.6010 |
| 3.5542 | 4.79 | 75 | 3.5621 |
| 3.5309 | 6.38 | 100 | 3.5117 |
| 3.496 | 7.98 | 125 | 3.4615 |
| 3.446 | 9.57 | 150 | 3.4173 |
| 3.34 | 11.17 | 175 | 3.3699 |
| 3.3581 | 12.77 | 200 | 3.3214 |
| 3.3136 | 14.36 | 225 | 3.2743 |
| 3.214 | 15.96 | 250 | 3.2227 |
| 3.2098 | 17.55 | 275 | 3.1738 |
| 3.1348 | 19.15 | 300 | 3.1153 |
| 3.0931 | 20.74 | 325 | 3.0561 |
| 3.0383 | 22.34 | 350 | 2.9922 |
| 2.9739 | 23.94 | 375 | 2.9269 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.0
- Tokenizers 0.15.0