File size: 3,450 Bytes
7e8ece8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e2028b6
7e8ece8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e2028b6
 
7e8ece8
 
e2028b6
7e8ece8
 
 
e2028b6
7e8ece8
 
 
 
e2028b6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7e8ece8
 
 
 
e2028b6
 
 
 
 
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
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
---
license: other
library_name: peft
tags:
- generated_from_trainer
base_model: beomi/gemma-ko-2b
model-index:
- name: gemma2_on_korean_summary
  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. -->

# gemma2_on_korean_summary

This model is a fine-tuned version of [beomi/gemma-ko-2b](https://huggingface.co/beomi/gemma-ko-2b) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9622

## 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-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 5
- total_train_batch_size: 10
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- training_steps: 800
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.6606        | 0.26  | 20   | 1.5454          |
| 1.4381        | 0.53  | 40   | 1.3247          |
| 1.2548        | 0.79  | 60   | 1.1921          |
| 1.1574        | 1.05  | 80   | 1.1227          |
| 1.0968        | 1.32  | 100  | 1.0727          |
| 1.0485        | 1.58  | 120  | 1.0316          |
| 1.0258        | 1.84  | 140  | 1.0019          |
| 0.9582        | 2.11  | 160  | 0.9785          |
| 0.906         | 2.37  | 180  | 0.9575          |
| 0.8837        | 2.63  | 200  | 0.9429          |
| 0.8763        | 2.89  | 220  | 0.9247          |
| 0.8295        | 3.16  | 240  | 0.9213          |
| 0.7799        | 3.42  | 260  | 0.9122          |
| 0.7742        | 3.68  | 280  | 0.8992          |
| 0.7708        | 3.95  | 300  | 0.8918          |
| 0.7196        | 4.21  | 320  | 0.8952          |
| 0.6908        | 4.47  | 340  | 0.8917          |
| 0.6977        | 4.74  | 360  | 0.8841          |
| 0.6789        | 5.0   | 380  | 0.8764          |
| 0.6198        | 5.26  | 400  | 0.9003          |
| 0.6203        | 5.53  | 420  | 0.9030          |
| 0.6169        | 5.79  | 440  | 0.8913          |
| 0.6111        | 6.05  | 460  | 0.8935          |
| 0.564         | 6.32  | 480  | 0.9096          |
| 0.5819        | 6.58  | 500  | 0.9027          |
| 0.5673        | 6.84  | 520  | 0.8997          |
| 0.5382        | 7.11  | 540  | 0.9181          |
| 0.5228        | 7.37  | 560  | 0.9197          |
| 0.5197        | 7.63  | 580  | 0.9254          |
| 0.5319        | 7.89  | 600  | 0.9158          |
| 0.4954        | 8.16  | 620  | 0.9394          |
| 0.4954        | 8.42  | 640  | 0.9345          |
| 0.4719        | 8.68  | 660  | 0.9388          |
| 0.4866        | 8.95  | 680  | 0.9390          |
| 0.4739        | 9.21  | 700  | 0.9528          |
| 0.4591        | 9.47  | 720  | 0.9534          |
| 0.4478        | 9.74  | 740  | 0.9542          |
| 0.4547        | 10.0  | 760  | 0.9539          |
| 0.4379        | 10.26 | 780  | 0.9617          |
| 0.4348        | 10.53 | 800  | 0.9622          |


### Framework versions

- PEFT 0.8.2
- Transformers 4.38.2
- Pytorch 2.1.2+cu121
- Datasets 2.17.0
- Tokenizers 0.15.0