File size: 2,025 Bytes
13609d3 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 |
---
library_name: peft
license: mit
base_model: unsloth/Phi-3.5-mini-instruct
tags:
- trl
- sft
- unsloth
- generated_from_trainer
model-index:
- name: Phi-3.5-mini-instruct-2024-10-28_15-54-04
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# Phi-3.5-mini-instruct-2024-10-28_15-54-04
This model is a fine-tuned version of [unsloth/Phi-3.5-mini-instruct](https://huggingface.co/unsloth/Phi-3.5-mini-instruct) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6209
## 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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.05
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 0.6797 | 0.2492 | 148 | 0.6636 |
| 0.6457 | 0.4983 | 296 | 0.6456 |
| 0.6394 | 0.7475 | 444 | 0.6378 |
| 0.6539 | 0.9966 | 592 | 0.6329 |
| 0.6116 | 1.2458 | 740 | 0.6299 |
| 0.617 | 1.4949 | 888 | 0.6284 |
| 0.5936 | 1.7441 | 1036 | 0.6254 |
| 0.5994 | 1.9933 | 1184 | 0.6231 |
| 0.6277 | 2.2424 | 1332 | 0.6226 |
| 0.6123 | 2.4916 | 1480 | 0.6217 |
| 0.6583 | 2.7407 | 1628 | 0.6210 |
| 0.5918 | 2.9899 | 1776 | 0.6209 |
### Framework versions
- PEFT 0.13.2
- Transformers 4.45.2
- Pytorch 2.4.0+cu121
- Datasets 2.19.2
- Tokenizers 0.20.1 |