File size: 3,446 Bytes
f21cbc1 | 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 75 76 77 78 79 80 81 82 83 | ---
license: mit
base_model: microsoft/phi-2
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: phi-2-coedit
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-2-coedit
This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7388
- Rouge1: 0.5206
- Rouge2: 0.4123
- Rougel: 0.4979
- Rougelsum: 0.5032
- Sacreblue: 28.1346
- Memory Used: 81917.5
- Cuda Allocated: 10795.7861
- Cuda Reserved: 74746.0
- Ram Usage: 24042.6719
- Em: 0.0
- Gen Len: 120.6545
## 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: 2e-05
- train_batch_size: 35
- eval_batch_size: 35
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 140
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Sacreblue | Memory Used | Cuda Allocated | Cuda Reserved | Ram Usage | Em | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:---------:|:-----------:|:--------------:|:-------------:|:----------:|:---:|:--------:|
| 0.5716 | 0.22 | 100 | 0.7558 | 0.5041 | 0.3927 | 0.4809 | 0.4853 | 26.9798 | 81917.5 | 10795.811 | 74738.0 | 22888.4102 | 0.0 | 120.3347 |
| 0.5407 | 0.44 | 200 | 0.7404 | 0.5241 | 0.4171 | 0.5013 | 0.5068 | 27.6806 | 81917.5 | 10795.814 | 74738.0 | 23733.9805 | 0.0 | 120.8277 |
| 0.5324 | 0.66 | 300 | 0.7230 | 0.5176 | 0.4093 | 0.4947 | 0.5002 | 27.5145 | 81917.5 | 10795.8184 | 74738.0 | 23831.1484 | 0.0 | 120.576 |
| 0.5107 | 0.88 | 400 | 0.7161 | 0.5256 | 0.4167 | 0.5042 | 0.5092 | 28.1274 | 81917.5 | 10795.7935 | 74738.0 | 23891.7891 | 0.0 | 120.5225 |
| 0.4374 | 1.1 | 500 | 0.7495 | 0.5237 | 0.414 | 0.501 | 0.5059 | 28.0405 | 81917.5 | 10795.7861 | 74746.0 | 23922.043 | 0.0 | 120.3181 |
| 0.3515 | 1.32 | 600 | 0.7418 | 0.5216 | 0.4133 | 0.499 | 0.5049 | 28.0528 | 81917.5 | 10795.7832 | 74746.0 | 23973.8164 | 0.0 | 120.6453 |
| 0.3449 | 1.54 | 700 | 0.7386 | 0.5242 | 0.4163 | 0.5016 | 0.5075 | 28.3145 | 81917.5 | 10795.8066 | 74746.0 | 23950.1016 | 0.0 | 120.5367 |
| 0.3375 | 1.76 | 800 | 0.7354 | 0.5194 | 0.4124 | 0.4973 | 0.5025 | 28.0252 | 81917.5 | 10795.814 | 74746.0 | 23931.0 | 0.0 | 120.6476 |
| 0.3373 | 1.98 | 900 | 0.7388 | 0.5206 | 0.4123 | 0.4979 | 0.5032 | 28.1346 | 81917.5 | 10795.7861 | 74746.0 | 24042.6719 | 0.0 | 120.6545 |
### Framework versions
- Transformers 4.39.3
- Pytorch 2.2.2
- Datasets 2.18.0
- Tokenizers 0.15.2
|