File size: 3,309 Bytes
252b216
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: transformers
license: apache-2.0
base_model: tsavage68/IE_M2_1000steps_1e7rate_SFT
tags:
- trl
- dpo
- generated_from_trainer
model-index:
- name: IE_M2_350steps_1e8rate_05beta_cSFTDPO
  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. -->

# IE_M2_350steps_1e8rate_05beta_cSFTDPO

This model is a fine-tuned version of [tsavage68/IE_M2_1000steps_1e7rate_SFT](https://huggingface.co/tsavage68/IE_M2_1000steps_1e7rate_SFT) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6608
- Rewards/chosen: 0.0105
- Rewards/rejected: -0.0599
- Rewards/accuracies: 0.3850
- Rewards/margins: 0.0703
- Logps/rejected: -41.1416
- Logps/chosen: -42.1846
- Logits/rejected: -2.9156
- Logits/chosen: -2.8543

## 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-08
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 350

### Training results

| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 0.6945        | 0.4   | 50   | 0.6933          | 0.0121         | 0.0098           | 0.2450             | 0.0023          | -41.0022       | -42.1813     | -2.9159         | -2.8545       |
| 0.6936        | 0.8   | 100  | 0.6888          | 0.0052         | -0.0069          | 0.2150             | 0.0121          | -41.0356       | -42.1952     | -2.9158         | -2.8545       |
| 0.6822        | 1.2   | 150  | 0.6646          | 0.0021         | -0.0599          | 0.3650             | 0.0621          | -41.1417       | -42.2012     | -2.9157         | -2.8544       |
| 0.6637        | 1.6   | 200  | 0.6652          | 0.0023         | -0.0586          | 0.3600             | 0.0609          | -41.1390       | -42.2010     | -2.9157         | -2.8544       |
| 0.6647        | 2.0   | 250  | 0.6601          | 0.0043         | -0.0670          | 0.3900             | 0.0713          | -41.1557       | -42.1968     | -2.9157         | -2.8544       |
| 0.6697        | 2.4   | 300  | 0.6624          | 0.0067         | -0.0606          | 0.3800             | 0.0673          | -41.1430       | -42.1922     | -2.9156         | -2.8543       |
| 0.6579        | 2.8   | 350  | 0.6608          | 0.0105         | -0.0599          | 0.3850             | 0.0703          | -41.1416       | -42.1846     | -2.9156         | -2.8543       |


### Framework versions

- Transformers 4.44.2
- Pytorch 2.0.0+cu117
- Datasets 3.0.0
- Tokenizers 0.19.1