File size: 4,525 Bytes
f54f344
 
 
1e38cb4
f54f344
 
 
 
 
 
 
 
 
 
 
 
 
1e38cb4
f54f344
1e38cb4
 
f54f344
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
103
104
105
106
107
108
109
110
111
112
---
license: apache-2.0
tags:
- image-classification
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: new_exper3
  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. -->

# new_exper3

This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the sudo-s/herbier_mesuem1 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3000
- Accuracy: 0.9298

## 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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
- mixed_precision_training: Apex, opt level O1

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 4.093         | 0.16  | 100  | 4.1045          | 0.1885   |
| 3.5057        | 0.31  | 200  | 3.4448          | 0.3231   |
| 2.9116        | 0.47  | 300  | 2.9483          | 0.4537   |
| 2.561         | 0.63  | 400  | 2.5700          | 0.5258   |
| 2.1611        | 0.78  | 500  | 2.1721          | 0.6145   |
| 1.715         | 0.94  | 600  | 1.8255          | 0.6407   |
| 1.2752        | 1.1   | 700  | 1.5340          | 0.7051   |
| 1.2487        | 1.25  | 800  | 1.3533          | 0.7201   |
| 1.0333        | 1.41  | 900  | 1.1474          | 0.7826   |
| 0.8856        | 1.56  | 1000 | 1.0914          | 0.7645   |
| 0.7512        | 1.72  | 1100 | 0.8893          | 0.8119   |
| 0.747         | 1.88  | 1200 | 0.8370          | 0.8304   |
| 0.5082        | 2.03  | 1300 | 0.7131          | 0.8566   |
| 0.4449        | 2.19  | 1400 | 0.6573          | 0.8547   |
| 0.2912        | 2.35  | 1500 | 0.6184          | 0.8597   |
| 0.285         | 2.5   | 1600 | 0.5974          | 0.8570   |
| 0.2267        | 2.66  | 1700 | 0.5621          | 0.8647   |
| 0.2553        | 2.82  | 1800 | 0.5044          | 0.8816   |
| 0.2029        | 2.97  | 1900 | 0.4342          | 0.8955   |
| 0.1763        | 3.13  | 2000 | 0.4487          | 0.8905   |
| 0.1418        | 3.29  | 2100 | 0.4173          | 0.9005   |
| 0.0563        | 3.44  | 2200 | 0.3870          | 0.9048   |
| 0.0579        | 3.6   | 2300 | 0.3849          | 0.9036   |
| 0.166         | 3.76  | 2400 | 0.3933          | 0.9025   |
| 0.11          | 3.91  | 2500 | 0.3918          | 0.9056   |
| 0.0356        | 4.07  | 2600 | 0.3298          | 0.9202   |
| 0.0513        | 4.23  | 2700 | 0.3371          | 0.9210   |
| 0.0762        | 4.38  | 2800 | 0.3253          | 0.9225   |
| 0.018         | 4.54  | 2900 | 0.3467          | 0.9148   |
| 0.0263        | 4.69  | 3000 | 0.3544          | 0.9144   |
| 0.0205        | 4.85  | 3100 | 0.3340          | 0.9221   |
| 0.0237        | 5.01  | 3200 | 0.3353          | 0.9144   |
| 0.013         | 5.16  | 3300 | 0.3218          | 0.9229   |
| 0.0116        | 5.32  | 3400 | 0.3088          | 0.9291   |
| 0.0119        | 5.48  | 3500 | 0.3047          | 0.9279   |
| 0.0098        | 5.63  | 3600 | 0.3063          | 0.9283   |
| 0.0086        | 5.79  | 3700 | 0.3074          | 0.9268   |
| 0.0081        | 5.95  | 3800 | 0.3220          | 0.9237   |
| 0.0078        | 6.1   | 3900 | 0.3064          | 0.9268   |
| 0.0074        | 6.26  | 4000 | 0.3062          | 0.9279   |
| 0.0068        | 6.42  | 4100 | 0.3051          | 0.9291   |
| 0.006         | 6.57  | 4200 | 0.3000          | 0.9298   |
| 0.0075        | 6.73  | 4300 | 0.3010          | 0.9310   |
| 0.0057        | 6.89  | 4400 | 0.3037          | 0.9298   |
| 0.0058        | 7.04  | 4500 | 0.3071          | 0.9279   |
| 0.0075        | 7.2   | 4600 | 0.3075          | 0.9283   |
| 0.0066        | 7.36  | 4700 | 0.3077          | 0.9295   |
| 0.0056        | 7.51  | 4800 | 0.3084          | 0.9295   |
| 0.0053        | 7.67  | 4900 | 0.3064          | 0.9310   |
| 0.0057        | 7.82  | 5000 | 0.3068          | 0.9318   |
| 0.0055        | 7.98  | 5100 | 0.3068          | 0.9318   |


### Framework versions

- Transformers 4.19.4
- Pytorch 1.5.1
- Datasets 2.3.2
- Tokenizers 0.12.1