File size: 3,689 Bytes
b6e210a |
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
base_model: google/gemma-2b
metrics:
- accuracy
model-index:
- name: RM-HH-Gemma_helpful_human_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-Gemma_helpful_human_loraR64_20000_gemma2b_shuffleTrue_extractchosenFalse
This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co/google/gemma-2b) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6198
- Accuracy: 0.6540
## 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 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.7745 | 0.06 | 250 | 0.7362 | 0.5088 |
| 0.6966 | 0.11 | 500 | 0.7087 | 0.5498 |
| 0.6929 | 0.17 | 750 | 0.6929 | 0.5814 |
| 0.702 | 0.22 | 1000 | 0.6814 | 0.5939 |
| 0.6633 | 0.28 | 1250 | 0.6735 | 0.6049 |
| 0.6529 | 0.33 | 1500 | 0.6669 | 0.6094 |
| 0.6487 | 0.39 | 1750 | 0.6610 | 0.6189 |
| 0.6737 | 0.45 | 2000 | 0.6536 | 0.6254 |
| 0.6314 | 0.5 | 2250 | 0.6501 | 0.6269 |
| 0.6474 | 0.56 | 2500 | 0.6454 | 0.6304 |
| 0.6225 | 0.61 | 2750 | 0.6429 | 0.6335 |
| 0.6338 | 0.67 | 3000 | 0.6393 | 0.6360 |
| 0.6268 | 0.72 | 3250 | 0.6360 | 0.6400 |
| 0.633 | 0.78 | 3500 | 0.6346 | 0.6425 |
| 0.641 | 0.83 | 3750 | 0.6305 | 0.6440 |
| 0.6439 | 0.89 | 4000 | 0.6286 | 0.6470 |
| 0.6123 | 0.95 | 4250 | 0.6274 | 0.6475 |
| 0.6082 | 1.0 | 4500 | 0.6277 | 0.6535 |
| 0.6275 | 1.06 | 4750 | 0.6267 | 0.6540 |
| 0.589 | 1.11 | 5000 | 0.6276 | 0.6535 |
| 0.588 | 1.17 | 5250 | 0.6297 | 0.6550 |
| 0.6126 | 1.22 | 5500 | 0.6305 | 0.6535 |
| 0.6216 | 1.28 | 5750 | 0.6286 | 0.6525 |
| 0.6071 | 1.34 | 6000 | 0.6269 | 0.6515 |
| 0.6063 | 1.39 | 6250 | 0.6271 | 0.6505 |
| 0.6166 | 1.45 | 6500 | 0.6246 | 0.6525 |
| 0.6076 | 1.5 | 6750 | 0.6230 | 0.6565 |
| 0.6007 | 1.56 | 7000 | 0.6233 | 0.6545 |
| 0.6452 | 1.61 | 7250 | 0.6205 | 0.6540 |
| 0.5932 | 1.67 | 7500 | 0.6207 | 0.6535 |
| 0.6093 | 1.72 | 7750 | 0.6207 | 0.6530 |
| 0.6183 | 1.78 | 8000 | 0.6206 | 0.6535 |
| 0.6244 | 1.84 | 8250 | 0.6200 | 0.6545 |
| 0.6183 | 1.89 | 8500 | 0.6199 | 0.6545 |
| 0.6281 | 1.95 | 8750 | 0.6198 | 0.6540 |
### Framework versions
- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2 |