File size: 3,671 Bytes
d765114 |
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 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 |
---
license: gemma
library_name: peft
tags:
- trl
- reward-trainer
- generated_from_trainer
metrics:
- accuracy
base_model: google/gemma-2b
model-index:
- name: RM-HH-Gemma_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenTrue
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. -->
# RM-HH-Gemma_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenTrue
This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co./google/gemma-2b) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0495
- Accuracy: 0.9820
## 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: 1.41e-05
- train_batch_size: 1
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.9068 | 0.03 | 250 | 0.5546 | 0.7177 |
| 0.5566 | 0.06 | 500 | 0.2048 | 0.9170 |
| 0.5143 | 0.08 | 750 | 0.1646 | 0.9370 |
| 0.4865 | 0.11 | 1000 | 0.1396 | 0.9457 |
| 0.4771 | 0.14 | 1250 | 0.1204 | 0.9510 |
| 0.4452 | 0.17 | 1500 | 0.1118 | 0.9565 |
| 0.436 | 0.19 | 1750 | 0.1063 | 0.9570 |
| 0.4433 | 0.22 | 2000 | 0.0942 | 0.9615 |
| 0.4541 | 0.25 | 2250 | 0.0878 | 0.9647 |
| 0.4361 | 0.28 | 2500 | 0.0822 | 0.9672 |
| 0.4626 | 0.31 | 2750 | 0.0766 | 0.9700 |
| 0.4595 | 0.33 | 3000 | 0.0714 | 0.9720 |
| 0.4375 | 0.36 | 3250 | 0.0720 | 0.9715 |
| 0.4338 | 0.39 | 3500 | 0.0693 | 0.9727 |
| 0.4082 | 0.42 | 3750 | 0.0675 | 0.9720 |
| 0.4306 | 0.44 | 4000 | 0.0635 | 0.9745 |
| 0.4296 | 0.47 | 4250 | 0.0629 | 0.9750 |
| 0.4318 | 0.5 | 4500 | 0.0590 | 0.9767 |
| 0.4226 | 0.53 | 4750 | 0.0575 | 0.9775 |
| 0.435 | 0.56 | 5000 | 0.0556 | 0.9785 |
| 0.4501 | 0.58 | 5250 | 0.0557 | 0.9790 |
| 0.3923 | 0.61 | 5500 | 0.0542 | 0.9785 |
| 0.4222 | 0.64 | 5750 | 0.0541 | 0.9790 |
| 0.3891 | 0.67 | 6000 | 0.0538 | 0.9787 |
| 0.4123 | 0.69 | 6250 | 0.0551 | 0.9790 |
| 0.3805 | 0.72 | 6500 | 0.0521 | 0.9805 |
| 0.4269 | 0.75 | 6750 | 0.0529 | 0.9800 |
| 0.382 | 0.78 | 7000 | 0.0530 | 0.9802 |
| 0.422 | 0.81 | 7250 | 0.0517 | 0.9812 |
| 0.4621 | 0.83 | 7500 | 0.0506 | 0.9812 |
| 0.3963 | 0.86 | 7750 | 0.0498 | 0.9820 |
| 0.4097 | 0.89 | 8000 | 0.0495 | 0.9820 |
| 0.4705 | 0.92 | 8250 | 0.0492 | 0.9822 |
| 0.4248 | 0.94 | 8500 | 0.0493 | 0.9820 |
| 0.3938 | 0.97 | 8750 | 0.0495 | 0.9820 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2 |