File size: 2,963 Bytes
b0907b2
1b13eca
 
639ba6f
 
1b13eca
 
 
 
 
 
 
639ba6f
1b13eca
 
 
b0907b2
1b13eca
 
 
 
 
 
 
 
 
 
7339d99
1b13eca
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9c6d908
1b13eca
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
language:
- en
license:
- creativeml-openrail-m
tags:
- generated_from_trainer
- text generation
- pytorch
- casual-lm
metrics:
- accuracy
base_model: EleutherAI/pythia-125m-deduped
model-index:
- name: openchatgpt-neox-r1
  results: []
---

# openchatgpt-neox-r1

This model is a fine-tuned version of [EleutherAI/pythia-125m-deduped](https://huggingface.co/EleutherAI/pythia-125m-deduped) on the openchatgpt safe-r1 dataset.
It achieves the following results on the evaluation set:
- Loss: 1.3585
- Accuracy: 0.9169

## Model description

Finetune based on the inner workings of ChatGPT. I won't elaborate on that. You must have a faint idea of how prompt is made for it to be effective.

This is effectively a schizophrenic idea that met the light of day. Practically a collab of 3 students in a virtual shed.

BTW, Pythia is so much better omg.

## Intended uses & limitations

Intended uses & limitations fall in line with OpenAI's. Dataset used consists of safe texts (i.e. not highly sexual/erotica type stuff). NSFW version of the dataset is not planned to exist at the moment.

Keep in mind that this is a 125m version of GPT-NeoX (Pythia). My 1050Ti Mobile couldn't even handle that without gradient thingmabobs, 8BitAdam was also used. If anyone knows how to effectively finetune larger models on free colabs - feel free to let me know. Pile tokenizer also has one downside compared to native GPT-2/3 - `Assistant` is not 1 token, but 2.

## Training and evaluation data

Data was split in ratio of 95%/5%. Preproccess included removing mentions of OpenAI wherever it was not deemed appropriete (GPT-2 has one of the appropriete mentions). Whole dataset consists of just shy off 3k input-output pairs. One input has multiple outputs (read as: one message has multiple variants of an answer). <<<1% (3 total) are curated lines (i.e. a huge mistake was spotted that needed corrections). At least 3 lines (<<<1% of line count, but more of byte count) are broken.

Heavy bias on IT.

## Training procedure

Input and output were straight up concatenated due to the nature of how ChatGPT works.

EOS was being used after the final separator.

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 2
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 1.1311        | 1.0   | 1377 | 1.3116          | 0.9127   |
| 0.6691        | 2.0   | 2754 | 1.2978          | 0.9160   |
| 0.3463        | 3.0   | 4131 | 1.3585          | 0.9169   |


### Framework versions

- Transformers 4.25.1
- Pytorch 1.13.1+cu116
- Datasets 2.8.0
- Tokenizers 0.13.2