metadata
base_model: microsoft/Phi-3-mini-4k-instruct
library_name: peft
license: mit
tags:
- generated_from_trainer
model-index:
- name: phi3-mini-4k-adapter_4
results: []
phi3-mini-4k-adapter_4
This model is a fine-tuned version of microsoft/Phi-3-mini-4k-instruct on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.1769
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: 0.0002
- train_batch_size: 6
- eval_batch_size: 6
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 24
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.05
- training_steps: 200
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 7.4868 | 0.9091 | 10 | 5.1948 |
| 0.2094 | 1.8182 | 20 | 0.2065 |
| 0.1037 | 2.7273 | 30 | 0.1536 |
| 0.0781 | 3.6364 | 40 | 0.1529 |
| 0.0635 | 4.5455 | 50 | 0.1555 |
| 0.0612 | 5.4545 | 60 | 0.1575 |
| 0.0548 | 6.3636 | 70 | 0.1723 |
| 0.0502 | 7.2727 | 80 | 0.1769 |
Framework versions
- PEFT 0.11.1
- Transformers 4.43.1
- Pytorch 2.4.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1