File size: 3,676 Bytes
b6cf0f6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
be46998
b6cf0f6
be46998
 
 
 
 
 
 
 
 
b6cf0f6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
be46998
 
 
 
 
 
 
 
 
 
 
b6cf0f6
 
 
 
 
 
 
 
 
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
---
license: mit
library_name: peft
tags:
- trl
- dpo
- generated_from_trainer
base_model: microsoft/phi-2
model-index:
- name: phi-2-gpo-ultrafeedback-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. -->

# phi-2-gpo-ultrafeedback-lora

This model is a fine-tuned version of [microsoft/phi-2](https://huggingface.co/microsoft/phi-2) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0004
- Rewards/chosen: -0.0084
- Rewards/rejected: -0.0177
- Rewards/accuracies: 0.6700
- Rewards/margins: 0.0093
- Logps/rejected: -233.2047
- Logps/chosen: -261.0818
- Logits/rejected: 0.8824
- Logits/chosen: 0.7796

## 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: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 2

### Training results

| Training Loss | Epoch | Step | Logits/chosen | Logits/rejected | Logps/chosen | Logps/rejected | Validation Loss | Rewards/accuracies | Rewards/chosen | Rewards/margins | Rewards/rejected |
|:-------------:|:-----:|:----:|:-------------:|:---------------:|:------------:|:--------------:|:---------------:|:------------------:|:--------------:|:---------------:|:----------------:|
| 0.0026        | 0.21  | 100  | 0.8151        | 0.9175          | -260.2373    | -231.4896      | 0.0025          | 0.5080             | 0.0001         | 0.0006          | -0.0005          |
| 0.0023        | 0.42  | 200  | 0.8092        | 0.9120          | -260.3932    | -232.1152      | 0.0023          | 0.6560             | -0.0015        | 0.0053          | -0.0068          |
| 0.0022        | 0.63  | 300  | 0.7992        | 0.9022          | -260.9179    | -232.8447      | 0.0022          | 0.6700             | -0.0067        | 0.0073          | -0.0141          |
| 0.0021        | 0.84  | 400  | 0.7884        | 0.8914          | -261.1620    | -233.2157      | 0.0022          | 0.6640             | -0.0092        | 0.0086          | -0.0178          |
| 0.0022        | 1.05  | 500  | 0.7821        | 0.8853          | -261.1852    | -233.3614      | 0.0021          | 0.7100             | -0.0094        | 0.0098          | -0.0193          |
| 0.002         | 1.26  | 600  | 0.7815        | 0.8840          | -261.1207    | -233.2843      | 0.0021          | 0.6940             | -0.0088        | 0.0097          | -0.0185          |
| 0.0021        | 1.47  | 700  | 0.7790        | 0.8816          | -261.0788    | -233.2560      | 0.0021          | 0.7000             | -0.0083        | 0.0099          | -0.0182          |
| 0.0021        | 1.67  | 800  | 0.7781        | 0.8811          | -261.0643    | -233.2740      | 0.0021          | 0.6940             | -0.0082        | 0.0102          | -0.0184          |
| 0.0021        | 1.88  | 900  | 0.7806        | 0.8833          | -261.0922    | -233.2118      | 0.0021          | 0.6900             | -0.0085        | 0.0093          | -0.0178          |


### Framework versions

- PEFT 0.7.1
- Transformers 4.36.2
- Pytorch 2.1.2+cu118
- Datasets 2.14.6
- Tokenizers 0.15.2