metadata
license: mit
base_model: microsoft/phi-2
tags:
- generated_from_trainer
model-index:
- name: V0410MP1
results: []
V0410MP1
This model is a fine-tuned version of microsoft/phi-2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.3412
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: 0.0003
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine_with_restarts
- lr_scheduler_warmup_steps: 20
- num_epochs: 2
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 5.3721 | 0.18 | 20 | 4.1461 |
| 2.7382 | 0.36 | 40 | 1.8394 |
| 1.3779 | 0.54 | 60 | 0.9405 |
| 0.8284 | 0.73 | 80 | 0.5808 |
| 0.5725 | 0.91 | 100 | 0.3968 |
| 0.4463 | 1.09 | 120 | 0.3413 |
| 0.4602 | 1.27 | 140 | 0.3413 |
| 0.454 | 1.45 | 160 | 0.3412 |
| 0.4517 | 1.63 | 180 | 0.3412 |
| 0.4543 | 1.81 | 200 | 0.3413 |
| 0.4638 | 1.99 | 220 | 0.3412 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.2+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1