File size: 2,130 Bytes
86b8432
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1854635
86b8432
1854635
 
86b8432
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
base_model: JackFram/llama-68m
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: ShareGPT_llama2_68M
  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. -->

# ShareGPT_llama2_68M

This model is a fine-tuned version of [JackFram/llama-68m](https://huggingface.co/JackFram/llama-68m) on the anon8231489123/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json dataset.
It achieves the following results on the evaluation set:
- Loss: 2.3592
- Accuracy: 0.5813

## 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: 0.0001
- train_batch_size: 24
- eval_batch_size: 48
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.05
- num_epochs: 3.0

### Training results

| Training Loss | Epoch  | Step  | Validation Loss | Accuracy |
|:-------------:|:------:|:-----:|:---------------:|:--------:|
| 3.406         | 0.2644 | 1000  | 3.2345          | 0.5035   |
| 2.8119        | 0.5288 | 2000  | 2.8214          | 0.5365   |
| 2.607         | 0.7932 | 3000  | 2.6572          | 0.5498   |
| 2.4734        | 1.0576 | 4000  | 2.5736          | 0.5582   |
| 2.4335        | 1.3221 | 5000  | 2.5336          | 0.5617   |
| 2.3858        | 1.5865 | 6000  | 2.4896          | 0.5663   |
| 2.374         | 1.8509 | 7000  | 2.4711          | 0.5681   |
| 2.3044        | 2.1153 | 8000  | 2.4485          | 0.5702   |
| 2.2873        | 2.3797 | 9000  | 2.4132          | 0.5747   |
| 2.2669        | 2.6441 | 10000 | 2.3879          | 0.5778   |
| 2.2463        | 2.9085 | 11000 | 2.3636          | 0.5808   |


### Framework versions

- Transformers 4.41.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1