File size: 4,480 Bytes
2715002 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 |
---
license: gemma
library_name: peft
tags:
- trl
- reward-trainer
- generated_from_trainer
base_model: google/gemma-2b
metrics:
- accuracy
model-index:
- name: RM-HH-AllMix_helpful_gpt3_loraR64_20000_gemma2b_shuffleTrue_extractchosenFalse
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-AllMix_helpful_gpt3_loraR64_20000_gemma2b_shuffleTrue_extractchosenFalse
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: 0.5220
- Accuracy: 0.7437
## 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: 2.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.7063 | 0.04 | 250 | 0.6784 | 0.5939 |
| 0.6441 | 0.08 | 500 | 0.6032 | 0.6613 |
| 0.582 | 0.13 | 750 | 0.5617 | 0.6921 |
| 0.5045 | 0.17 | 1000 | 0.5495 | 0.6985 |
| 0.5345 | 0.21 | 1250 | 0.5444 | 0.7034 |
| 0.53 | 0.25 | 1500 | 0.5522 | 0.7076 |
| 0.5325 | 0.29 | 1750 | 0.5550 | 0.7061 |
| 0.5145 | 0.33 | 2000 | 0.5596 | 0.7121 |
| 0.5156 | 0.38 | 2250 | 0.5480 | 0.7143 |
| 0.4995 | 0.42 | 2500 | 0.5477 | 0.7181 |
| 0.5329 | 0.46 | 2750 | 0.5350 | 0.7207 |
| 0.5037 | 0.5 | 3000 | 0.5472 | 0.7196 |
| 0.5417 | 0.54 | 3250 | 0.5233 | 0.7249 |
| 0.5179 | 0.59 | 3500 | 0.5230 | 0.7256 |
| 0.5264 | 0.63 | 3750 | 0.5196 | 0.7286 |
| 0.4931 | 0.67 | 4000 | 0.5267 | 0.7279 |
| 0.5114 | 0.71 | 4250 | 0.5202 | 0.7317 |
| 0.4735 | 0.75 | 4500 | 0.5238 | 0.7332 |
| 0.4902 | 0.79 | 4750 | 0.5294 | 0.7332 |
| 0.5483 | 0.84 | 5000 | 0.5165 | 0.7343 |
| 0.548 | 0.88 | 5250 | 0.5070 | 0.7350 |
| 0.4918 | 0.92 | 5500 | 0.5115 | 0.7384 |
| 0.5079 | 0.96 | 5750 | 0.5108 | 0.7369 |
| 0.49 | 1.0 | 6000 | 0.5127 | 0.7388 |
| 0.5161 | 1.05 | 6250 | 0.5103 | 0.7392 |
| 0.4573 | 1.09 | 6500 | 0.5226 | 0.7369 |
| 0.4973 | 1.13 | 6750 | 0.5208 | 0.7358 |
| 0.5163 | 1.17 | 7000 | 0.5135 | 0.7373 |
| 0.4857 | 1.21 | 7250 | 0.5188 | 0.7381 |
| 0.4996 | 1.25 | 7500 | 0.5200 | 0.7384 |
| 0.5029 | 1.3 | 7750 | 0.5185 | 0.7388 |
| 0.4983 | 1.34 | 8000 | 0.5177 | 0.7384 |
| 0.4718 | 1.38 | 8250 | 0.5186 | 0.7392 |
| 0.4723 | 1.42 | 8500 | 0.5204 | 0.7381 |
| 0.5238 | 1.46 | 8750 | 0.5143 | 0.7403 |
| 0.4613 | 1.51 | 9000 | 0.5178 | 0.7384 |
| 0.517 | 1.55 | 9250 | 0.5212 | 0.7377 |
| 0.495 | 1.59 | 9500 | 0.5181 | 0.7407 |
| 0.4865 | 1.63 | 9750 | 0.5191 | 0.7418 |
| 0.4799 | 1.67 | 10000 | 0.5231 | 0.7414 |
| 0.4546 | 1.71 | 10250 | 0.5241 | 0.7426 |
| 0.4673 | 1.76 | 10500 | 0.5256 | 0.7433 |
| 0.4598 | 1.8 | 10750 | 0.5259 | 0.7448 |
| 0.5035 | 1.84 | 11000 | 0.5245 | 0.7444 |
| 0.5113 | 1.88 | 11250 | 0.5236 | 0.7433 |
| 0.4821 | 1.92 | 11500 | 0.5230 | 0.7433 |
| 0.5071 | 1.97 | 11750 | 0.5220 | 0.7437 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2 |