File size: 3,668 Bytes
aefb0c8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
base_model: Qwen/Qwen2-0.5B-Instruct
library_name: peft
license: apache-2.0
tags:
- trl
- dpo
- generated_from_trainer
model-index:
- name: dpo-model-lora
  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. -->

# dpo-model-lora

This model is a fine-tuned version of [Qwen/Qwen2-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2-0.5B-Instruct) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6534
- Rewards/chosen: -0.7320
- Rewards/rejected: -0.8303
- Rewards/accuracies: 0.6172
- Rewards/margins: 0.0983
- Logps/rejected: -359.0921
- Logps/chosen: -378.4928
- Logits/rejected: -2.2715
- Logits/chosen: -2.3471

## 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: 5e-06
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- total_eval_batch_size: 64
- 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 | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 0.6884        | 0.1030 | 50   | 0.6879          | -0.0543        | -0.0734          | 0.6484             | 0.0191          | -351.5229      | -371.7161    | -2.2877         | -2.3628       |
| 0.6787        | 0.2060 | 100  | 0.6770          | -0.1811        | -0.2114          | 0.6016             | 0.0303          | -352.9030      | -372.9836    | -2.2815         | -2.3565       |
| 0.6721        | 0.3090 | 150  | 0.6721          | -0.2679        | -0.3094          | 0.6562             | 0.0415          | -353.8831      | -373.8524    | -2.2782         | -2.3536       |
| 0.6668        | 0.4119 | 200  | 0.6665          | -0.4037        | -0.4625          | 0.6016             | 0.0588          | -355.4139      | -375.2100    | -2.2758         | -2.3515       |
| 0.6597        | 0.5149 | 250  | 0.6612          | -0.4907        | -0.5505          | 0.6172             | 0.0598          | -356.2946      | -376.0805    | -2.2757         | -2.3510       |
| 0.6581        | 0.6179 | 300  | 0.6578          | -0.6137        | -0.6975          | 0.625              | 0.0838          | -357.7639      | -377.3098    | -2.2736         | -2.3491       |
| 0.6536        | 0.7209 | 350  | 0.6556          | -0.6458        | -0.7367          | 0.6328             | 0.0909          | -358.1565      | -377.6311    | -2.2732         | -2.3489       |
| 0.6486        | 0.8239 | 400  | 0.6556          | -0.7025        | -0.7958          | 0.6328             | 0.0933          | -358.7473      | -378.1981    | -2.2737         | -2.3493       |
| 0.649         | 0.9269 | 450  | 0.6556          | -0.7432        | -0.8327          | 0.6484             | 0.0896          | -359.1166      | -378.6048    | -2.2726         | -2.3482       |


### Framework versions

- PEFT 0.12.0
- Transformers 4.44.2
- Pytorch 2.4.0+cu121
- Datasets 2.21.0
- Tokenizers 0.19.1