File size: 2,373 Bytes
a33c5e7
 
 
 
 
aeebb44
a33c5e7
 
 
aeebb44
 
 
 
 
a33c5e7
 
 
 
 
 
 
 
 
 
aeebb44
a33c5e7
aeebb44
 
 
 
 
 
 
 
 
a33c5e7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: transformers
license: llama3
base_model: meta-llama/Meta-Llama-3-8B-Instruct
tags:
- alignment-handbook
- trl
- simpo
- generated_from_trainer
- trl
- simpo
- generated_from_trainer
datasets:
- yakazimir/llama3-ultrafeedback-armorm
model-index:
- name: llama3instruct_-orpo-10-0_5-1e-6-1_best
  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. -->

# llama3instruct_-orpo-10-0_5-1e-6-1_best

This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the yakazimir/llama3-ultrafeedback-armorm dataset.
It achieves the following results on the evaluation set:
- Loss: 2.4050
- Rewards/chosen: -11.3456
- Rewards/rejected: -15.4559
- Rewards/accuracies: 0.8102
- Rewards/margins: 4.1103
- Logps/rejected: -1.5456
- Logps/chosen: -1.1346
- Logits/rejected: -1.3837
- Logits/chosen: -1.4182

## 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: 1e-06
- train_batch_size: 2
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 16
- total_train_batch_size: 128
- 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: 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 |
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 2.1949        | 0.8743 | 400  | 2.4129          | -11.3049       | -15.3895         | 0.8133             | 4.0846          | -1.5389        | -1.1305      | -1.3507         | -1.3829       |


### Framework versions

- Transformers 4.44.2
- Pytorch 2.2.2+cu121
- Datasets 2.18.0
- Tokenizers 0.19.1