File size: 3,321 Bytes
d640c9c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: mit
base_model: indolem/indobert-base-uncased
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: sentiment-lora-r4a2d0.15-0
  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. -->

# sentiment-lora-r4a2d0.15-0

This model is a fine-tuned version of [indolem/indobert-base-uncased](https://huggingface.co/indolem/indobert-base-uncased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3513
- Accuracy: 0.8471
- Precision: 0.8147
- Recall: 0.8193
- F1: 0.8169

## 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: 30
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20.0

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1     |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| 0.5621        | 1.0   | 122  | 0.5100          | 0.7218   | 0.6593    | 0.6482 | 0.6527 |
| 0.5049        | 2.0   | 244  | 0.4890          | 0.7343   | 0.6945    | 0.7195 | 0.7011 |
| 0.4776        | 3.0   | 366  | 0.4480          | 0.7594   | 0.7150    | 0.7323 | 0.7216 |
| 0.4422        | 4.0   | 488  | 0.4104          | 0.7945   | 0.7524    | 0.7446 | 0.7482 |
| 0.4146        | 5.0   | 610  | 0.4257          | 0.7594   | 0.7202    | 0.7473 | 0.7283 |
| 0.3828        | 6.0   | 732  | 0.3869          | 0.8246   | 0.7880    | 0.7909 | 0.7894 |
| 0.3697        | 7.0   | 854  | 0.3959          | 0.8145   | 0.7766    | 0.7988 | 0.7854 |
| 0.3486        | 8.0   | 976  | 0.3808          | 0.8321   | 0.7961    | 0.8087 | 0.8018 |
| 0.3437        | 9.0   | 1098 | 0.3738          | 0.8271   | 0.7904    | 0.8001 | 0.7949 |
| 0.3317        | 10.0  | 1220 | 0.3643          | 0.8471   | 0.8159    | 0.8143 | 0.8151 |
| 0.3114        | 11.0  | 1342 | 0.3683          | 0.8271   | 0.7902    | 0.8051 | 0.7968 |
| 0.3035        | 12.0  | 1464 | 0.3660          | 0.8346   | 0.7988    | 0.8155 | 0.8061 |
| 0.3117        | 13.0  | 1586 | 0.3518          | 0.8471   | 0.8167    | 0.8118 | 0.8142 |
| 0.3048        | 14.0  | 1708 | 0.3533          | 0.8446   | 0.8115    | 0.8176 | 0.8144 |
| 0.2916        | 15.0  | 1830 | 0.3570          | 0.8421   | 0.8083    | 0.8158 | 0.8119 |
| 0.2832        | 16.0  | 1952 | 0.3579          | 0.8471   | 0.8138    | 0.8243 | 0.8187 |
| 0.284         | 17.0  | 2074 | 0.3496          | 0.8471   | 0.8153    | 0.8168 | 0.8160 |
| 0.2906        | 18.0  | 2196 | 0.3537          | 0.8446   | 0.8111    | 0.8201 | 0.8153 |
| 0.2805        | 19.0  | 2318 | 0.3505          | 0.8496   | 0.8186    | 0.8186 | 0.8186 |
| 0.2815        | 20.0  | 2440 | 0.3513          | 0.8471   | 0.8147    | 0.8193 | 0.8169 |


### Framework versions

- Transformers 4.39.3
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.15.2