metadata
license: mit
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
datasets:
- generator
base_model: microsoft/phi-2
model-index:
- name: phi-2_summarize
results: []
phi-2_summarize
This model is a fine-tuned version of microsoft/phi-2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 1.7266
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: 1
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 3
- training_steps: 400
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.9986 | 0.03 | 25 | 1.8320 |
| 1.8726 | 0.06 | 50 | 1.7854 |
| 1.7545 | 0.08 | 75 | 1.7727 |
| 1.8384 | 0.11 | 100 | 1.7614 |
| 1.7279 | 0.14 | 125 | 1.7561 |
| 1.7348 | 0.17 | 150 | 1.7503 |
| 1.8037 | 0.2 | 175 | 1.7442 |
| 1.7602 | 0.23 | 200 | 1.7418 |
| 1.8112 | 0.25 | 225 | 1.7372 |
| 1.7011 | 0.28 | 250 | 1.7339 |
| 1.6925 | 0.31 | 275 | 1.7312 |
| 1.6696 | 0.34 | 300 | 1.7295 |
| 1.7461 | 0.37 | 325 | 1.7275 |
| 1.6521 | 0.4 | 350 | 1.7270 |
| 1.6636 | 0.42 | 375 | 1.7267 |
| 1.7048 | 0.45 | 400 | 1.7266 |
Framework versions
- PEFT 0.10.0
- Transformers 4.39.3
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2