File size: 2,938 Bytes
1b0fa9f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
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_harmless_gpt3_20000_gemma2b_shuffleFalse_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_harmless_gpt3_20000_gemma2b_shuffleFalse_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: 0.0554
- Accuracy: 0.9722

## 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 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.6806        | 0.04  | 250  | 0.4207          | 0.8447   |
| 0.6465        | 0.08  | 500  | 0.2075          | 0.9485   |
| 0.5999        | 0.13  | 750  | 0.1133          | 0.9725   |
| 0.5725        | 0.17  | 1000 | 0.0804          | 0.9786   |
| 0.6223        | 0.21  | 1250 | 0.0783          | 0.9778   |
| 0.5595        | 0.25  | 1500 | 0.0632          | 0.9789   |
| 0.5956        | 0.29  | 1750 | 0.0589          | 0.9778   |
| 0.557         | 0.33  | 2000 | 0.0599          | 0.9756   |
| 0.5641        | 0.38  | 2250 | 0.0558          | 0.9767   |
| 0.5871        | 0.42  | 2500 | 0.0589          | 0.9744   |
| 0.5512        | 0.46  | 2750 | 0.0568          | 0.9741   |
| 0.5775        | 0.5   | 3000 | 0.0529          | 0.9756   |
| 0.5923        | 0.54  | 3250 | 0.0555          | 0.9748   |
| 0.548         | 0.59  | 3500 | 0.0577          | 0.9722   |
| 0.564         | 0.63  | 3750 | 0.0579          | 0.9722   |
| 0.563         | 0.67  | 4000 | 0.0599          | 0.9718   |
| 0.5932        | 0.71  | 4250 | 0.0561          | 0.9729   |
| 0.5247        | 0.75  | 4500 | 0.0569          | 0.9725   |
| 0.5472        | 0.79  | 4750 | 0.0579          | 0.9718   |
| 0.5704        | 0.84  | 5000 | 0.0556          | 0.9729   |
| 0.5456        | 0.88  | 5250 | 0.0550          | 0.9725   |
| 0.5563        | 0.92  | 5500 | 0.0545          | 0.9729   |
| 0.552         | 0.96  | 5750 | 0.0554          | 0.9722   |


### Framework versions

- PEFT 0.10.0
- Transformers 4.38.2
- Pytorch 2.1.2+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2