metadata
license: mit
base_model: microsoft/phi-1_5
tags:
- trl
- sft
- generated_from_trainer
datasets:
- generator
model-index:
- name: phi-1_5-sft-openhermes-v2
results: []
phi-1_5-sft-openhermes-v2
This model is a fine-tuned version of microsoft/phi-1_5 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 1.1750
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: 8e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 500
- num_epochs: 2
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.7865 | 0.0831 | 275 | 1.4033 |
| 1.3614 | 0.1663 | 550 | 1.3218 |
| 1.2986 | 0.2494 | 825 | 1.2788 |
| 1.2667 | 0.3325 | 1100 | 1.2531 |
| 1.2405 | 0.4157 | 1375 | 1.2376 |
| 1.2239 | 0.4988 | 1650 | 1.2237 |
| 1.2078 | 0.5819 | 1925 | 1.2122 |
| 1.2114 | 0.6651 | 2200 | 1.2005 |
| 1.2028 | 0.7482 | 2475 | 1.1915 |
| 1.173 | 0.8313 | 2750 | 1.1833 |
| 1.1782 | 0.9144 | 3025 | 1.1776 |
| 1.1805 | 0.9976 | 3300 | 1.1720 |
| 1.0112 | 1.0807 | 3575 | 1.1817 |
| 0.9988 | 1.1638 | 3850 | 1.1791 |
| 0.9919 | 1.2470 | 4125 | 1.1786 |
| 0.9886 | 1.3301 | 4400 | 1.1768 |
| 0.9904 | 1.4132 | 4675 | 1.1763 |
| 1.001 | 1.4964 | 4950 | 1.1756 |
| 0.9979 | 1.5795 | 5225 | 1.1751 |
| 0.9858 | 1.6626 | 5500 | 1.1750 |
| 0.9975 | 1.7458 | 5775 | 1.1750 |
| 0.9924 | 1.8289 | 6050 | 1.1750 |
| 0.9978 | 1.9120 | 6325 | 1.1750 |
| 0.9892 | 1.9952 | 6600 | 1.1750 |
Framework versions
- Transformers 4.42.4
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1