File size: 3,666 Bytes
1e1c504
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c5be93f
1e1c504
c5be93f
 
 
 
 
 
 
 
 
1e1c504
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c5be93f
1e1c504
 
 
c5be93f
 
 
 
 
 
 
 
 
 
 
1e1c504
 
 
 
 
 
 
 
 
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-dpo-ultrachat-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-dpo-ultrachat-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: 0.6872
- Rewards/chosen: -0.0312
- Rewards/rejected: -0.0436
- Rewards/accuracies: 0.3340
- Rewards/margins: 0.0124
- Logps/rejected: -98.5542
- Logps/chosen: -94.8435
- Logits/rejected: 0.7532
- Logits/chosen: 0.7326

## 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 |
|:-------------:|:-----:|:----:|:-------------:|:---------------:|:------------:|:--------------:|:---------------:|:------------------:|:--------------:|:---------------:|:----------------:|
| 0.693         | 0.21  | 100  | 0.7998        | 0.8176          | -91.7748     | -94.2804       | 0.6931          | 0.2680             | -0.0005        | 0.0004          | -0.0008          |
| 0.6922        | 0.42  | 200  | 0.7941        | 0.8121          | -91.9068     | -94.5141       | 0.6924          | 0.3020             | -0.0018        | 0.0014          | -0.0032          |
| 0.6917        | 0.63  | 300  | 0.7870        | 0.8057          | -92.2189     | -94.9659       | 0.6917          | 0.3100             | -0.0049        | 0.0028          | -0.0077          |
| 0.6905        | 0.84  | 400  | 0.7827        | 0.8012          | -92.4247     | -95.2509       | 0.6913          | 0.3280             | -0.0070        | 0.0036          | -0.0105          |
| 0.6898        | 1.05  | 500  | 0.6900        | -0.0142         | -0.0205      | 0.3360         | 0.0064          | -96.2490           | -93.1429       | 0.7903          | 0.7711           |
| 0.6882        | 1.26  | 600  | 0.6887        | -0.0217         | -0.0306      | 0.3340         | 0.0089          | -97.2594           | -93.8981       | 0.7722          | 0.7527           |
| 0.6858        | 1.47  | 700  | 0.6879        | -0.0274         | -0.0383      | 0.3280         | 0.0108          | -98.0249           | -94.4717       | 0.7600          | 0.7395           |
| 0.6857        | 1.67  | 800  | 0.6874        | -0.0303         | -0.0423      | 0.3340         | 0.0120          | -98.4270           | -94.7618       | 0.7548          | 0.7341           |
| 0.6866        | 1.88  | 900  | 0.6872        | -0.0313         | -0.0437      | 0.3420         | 0.0124          | -98.5655           | -94.8550       | 0.7528          | 0.7321           |


### Framework versions

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