File size: 4,520 Bytes
61d0bcd |
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: mit
library_name: peft
tags:
- trl
- reward-trainer
- generated_from_trainer
base_model: openai-community/gpt2-large
metrics:
- accuracy
model-index:
- name: RM-HH-AllMix_helpful_gpt3_loraR64_20000_gpt2-large_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_gpt2-large_shuffleTrue_extractchosenFalse
This model is a fine-tuned version of [openai-community/gpt2-large](https://huggingface.co./openai-community/gpt2-large) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4894
- Accuracy: 0.7351
## 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 |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.6628 | 0.04 | 250 | 0.6390 | 0.6277 |
| 0.598 | 0.08 | 500 | 0.5673 | 0.6933 |
| 0.5479 | 0.13 | 750 | 0.5415 | 0.7076 |
| 0.5397 | 0.17 | 1000 | 0.5308 | 0.7110 |
| 0.5094 | 0.21 | 1250 | 0.5261 | 0.7159 |
| 0.5142 | 0.25 | 1500 | 0.5203 | 0.7193 |
| 0.5414 | 0.29 | 1750 | 0.5161 | 0.7197 |
| 0.5189 | 0.33 | 2000 | 0.5131 | 0.7189 |
| 0.5151 | 0.38 | 2250 | 0.5100 | 0.7216 |
| 0.4942 | 0.42 | 2500 | 0.5089 | 0.7208 |
| 0.5067 | 0.46 | 2750 | 0.5057 | 0.7216 |
| 0.5026 | 0.5 | 3000 | 0.5041 | 0.7238 |
| 0.4926 | 0.54 | 3250 | 0.5038 | 0.7265 |
| 0.4931 | 0.59 | 3500 | 0.5022 | 0.7310 |
| 0.4946 | 0.63 | 3750 | 0.4993 | 0.7329 |
| 0.5058 | 0.67 | 4000 | 0.4968 | 0.7313 |
| 0.4822 | 0.71 | 4250 | 0.4963 | 0.7306 |
| 0.4924 | 0.75 | 4500 | 0.4961 | 0.7329 |
| 0.4654 | 0.8 | 4750 | 0.4959 | 0.7302 |
| 0.4924 | 0.84 | 5000 | 0.4971 | 0.7310 |
| 0.4674 | 0.88 | 5250 | 0.4948 | 0.7310 |
| 0.4704 | 0.92 | 5500 | 0.4950 | 0.7336 |
| 0.5089 | 0.96 | 5750 | 0.4905 | 0.7306 |
| 0.4673 | 1.0 | 6000 | 0.4929 | 0.7313 |
| 0.4594 | 1.05 | 6250 | 0.4932 | 0.7291 |
| 0.479 | 1.09 | 6500 | 0.4919 | 0.7332 |
| 0.5112 | 1.13 | 6750 | 0.4895 | 0.7355 |
| 0.4794 | 1.17 | 7000 | 0.4888 | 0.7332 |
| 0.5188 | 1.21 | 7250 | 0.4881 | 0.7340 |
| 0.4541 | 1.26 | 7500 | 0.4892 | 0.7359 |
| 0.4617 | 1.3 | 7750 | 0.4898 | 0.7366 |
| 0.4747 | 1.34 | 8000 | 0.4898 | 0.7362 |
| 0.4834 | 1.38 | 8250 | 0.4893 | 0.7389 |
| 0.4954 | 1.42 | 8500 | 0.4875 | 0.7385 |
| 0.5029 | 1.47 | 8750 | 0.4875 | 0.7385 |
| 0.4742 | 1.51 | 9000 | 0.4872 | 0.7400 |
| 0.4802 | 1.55 | 9250 | 0.4884 | 0.7393 |
| 0.5009 | 1.59 | 9500 | 0.4877 | 0.7400 |
| 0.4619 | 1.63 | 9750 | 0.4875 | 0.7396 |
| 0.4433 | 1.67 | 10000 | 0.4902 | 0.7404 |
| 0.4844 | 1.72 | 10250 | 0.4903 | 0.7400 |
| 0.4337 | 1.76 | 10500 | 0.4917 | 0.7400 |
| 0.4897 | 1.8 | 10750 | 0.4901 | 0.7396 |
| 0.4783 | 1.84 | 11000 | 0.4894 | 0.7366 |
| 0.4929 | 1.88 | 11250 | 0.4892 | 0.7359 |
| 0.4776 | 1.93 | 11500 | 0.4891 | 0.7362 |
| 0.4574 | 1.97 | 11750 | 0.4894 | 0.7351 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2 |