|
--- |
|
license: mit |
|
library_name: peft |
|
tags: |
|
- trl |
|
- sft |
|
- generated_from_trainer |
|
base_model: microsoft/phi-1_5 |
|
model-index: |
|
- name: phi-1_5-finetuned-question-generation |
|
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-1_5-finetuned-question-generation |
|
|
|
This model is a fine-tuned version of [microsoft/phi-1_5](https://huggingface.co/microsoft/phi-1_5) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.9170 |
|
|
|
## 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: 0.0001 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 2 |
|
- 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: linear |
|
- lr_scheduler_warmup_steps: 100 |
|
- num_epochs: 5 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 2.9836 | 0.09 | 100 | 2.8641 | |
|
| 2.8536 | 0.17 | 200 | 2.7929 | |
|
| 2.8051 | 0.26 | 300 | 2.7567 | |
|
| 2.7782 | 0.35 | 400 | 2.7092 | |
|
| 2.7542 | 0.44 | 500 | 2.6946 | |
|
| 2.6978 | 0.52 | 600 | 2.6719 | |
|
| 2.6833 | 0.61 | 700 | 2.6497 | |
|
| 2.6504 | 0.7 | 800 | 2.6172 | |
|
| 2.6228 | 0.78 | 900 | 2.6008 | |
|
| 2.6219 | 0.87 | 1000 | 2.5802 | |
|
| 2.5629 | 0.96 | 1100 | 2.5519 | |
|
| 2.5315 | 1.05 | 1200 | 2.5255 | |
|
| 2.4813 | 1.13 | 1300 | 2.5156 | |
|
| 2.4539 | 1.22 | 1400 | 2.4884 | |
|
| 2.4466 | 1.31 | 1500 | 2.4660 | |
|
| 2.4205 | 1.39 | 1600 | 2.4431 | |
|
| 2.3937 | 1.48 | 1700 | 2.4238 | |
|
| 2.3686 | 1.57 | 1800 | 2.4069 | |
|
| 2.3209 | 1.66 | 1900 | 2.3826 | |
|
| 2.3409 | 1.74 | 2000 | 2.3606 | |
|
| 2.2874 | 1.83 | 2100 | 2.3453 | |
|
| 2.309 | 1.92 | 2200 | 2.3222 | |
|
| 2.2676 | 2.01 | 2300 | 2.2981 | |
|
| 2.1734 | 2.09 | 2400 | 2.2892 | |
|
| 2.1495 | 2.18 | 2500 | 2.2549 | |
|
| 2.1163 | 2.27 | 2600 | 2.2401 | |
|
| 2.1 | 2.35 | 2700 | 2.2317 | |
|
| 2.1046 | 2.44 | 2800 | 2.2153 | |
|
| 2.1138 | 2.53 | 2900 | 2.1938 | |
|
| 2.0691 | 2.62 | 3000 | 2.1775 | |
|
| 2.0945 | 2.7 | 3100 | 2.1563 | |
|
| 2.045 | 2.79 | 3200 | 2.1408 | |
|
| 2.0212 | 2.88 | 3300 | 2.1229 | |
|
| 2.0011 | 2.96 | 3400 | 2.1156 | |
|
| 1.983 | 3.05 | 3500 | 2.0942 | |
|
| 1.9309 | 3.14 | 3600 | 2.0769 | |
|
| 1.8844 | 3.23 | 3700 | 2.0709 | |
|
| 1.9085 | 3.31 | 3800 | 2.0589 | |
|
| 1.8827 | 3.4 | 3900 | 2.0405 | |
|
| 1.8511 | 3.49 | 4000 | 2.0310 | |
|
| 1.8807 | 3.57 | 4100 | 2.0170 | |
|
| 1.8437 | 3.66 | 4200 | 2.0045 | |
|
| 1.8667 | 3.75 | 4300 | 2.0036 | |
|
| 1.8081 | 3.84 | 4400 | 1.9886 | |
|
| 1.8688 | 3.92 | 4500 | 1.9767 | |
|
| 1.8187 | 4.01 | 4600 | 1.9652 | |
|
| 1.7511 | 4.1 | 4700 | 1.9592 | |
|
| 1.7384 | 4.18 | 4800 | 1.9558 | |
|
| 1.7843 | 4.27 | 4900 | 1.9474 | |
|
| 1.7389 | 4.36 | 5000 | 1.9412 | |
|
| 1.7465 | 4.45 | 5100 | 1.9346 | |
|
| 1.7483 | 4.53 | 5200 | 1.9290 | |
|
| 1.7149 | 4.62 | 5300 | 1.9246 | |
|
| 1.7154 | 4.71 | 5400 | 1.9211 | |
|
| 1.7637 | 4.8 | 5500 | 1.9188 | |
|
| 1.7559 | 4.88 | 5600 | 1.9181 | |
|
| 1.7204 | 4.97 | 5700 | 1.9170 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.9.0 |
|
- Transformers 4.38.2 |
|
- Pytorch 2.1.2+cu121 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.15.2 |
