File size: 4,880 Bytes
ac1b466
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
113
114
115
116
117
118
119
120
121
122
123
124
125
126
---
license: apache-2.0
base_model: facebook/deit-small-patch16-224
tags:
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: hushem_40x_deit_small_adamax_00001_fold2
  results:
  - task:
      name: Image Classification
      type: image-classification
    dataset:
      name: imagefolder
      type: imagefolder
      config: default
      split: test
      args: default
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.7777777777777778
---

<!-- 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. -->

# hushem_40x_deit_small_adamax_00001_fold2

This model is a fine-tuned version of [facebook/deit-small-patch16-224](https://huggingface.co/facebook/deit-small-patch16-224) on the imagefolder dataset.
It achieves the following results on the evaluation set:
- Loss: 1.9649
- Accuracy: 0.7778

## 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: 1e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.2083        | 1.0   | 215   | 0.9013          | 0.7111   |
| 0.0287        | 2.0   | 430   | 0.8511          | 0.6889   |
| 0.0032        | 3.0   | 645   | 0.9572          | 0.7556   |
| 0.0007        | 4.0   | 860   | 1.0000          | 0.7778   |
| 0.0005        | 5.0   | 1075  | 1.0506          | 0.7778   |
| 0.0003        | 6.0   | 1290  | 1.0611          | 0.8      |
| 0.0002        | 7.0   | 1505  | 1.1165          | 0.8      |
| 0.0002        | 8.0   | 1720  | 1.1228          | 0.8      |
| 0.0001        | 9.0   | 1935  | 1.1485          | 0.8      |
| 0.0001        | 10.0  | 2150  | 1.1778          | 0.8      |
| 0.0001        | 11.0  | 2365  | 1.2118          | 0.8      |
| 0.0001        | 12.0  | 2580  | 1.2131          | 0.8      |
| 0.0001        | 13.0  | 2795  | 1.2712          | 0.8      |
| 0.0           | 14.0  | 3010  | 1.2931          | 0.8      |
| 0.0           | 15.0  | 3225  | 1.3148          | 0.8      |
| 0.0           | 16.0  | 3440  | 1.3419          | 0.8      |
| 0.0           | 17.0  | 3655  | 1.3678          | 0.8      |
| 0.0           | 18.0  | 3870  | 1.3854          | 0.8      |
| 0.0           | 19.0  | 4085  | 1.4044          | 0.8      |
| 0.0           | 20.0  | 4300  | 1.4309          | 0.8      |
| 0.0           | 21.0  | 4515  | 1.4555          | 0.8      |
| 0.0           | 22.0  | 4730  | 1.4779          | 0.8      |
| 0.0           | 23.0  | 4945  | 1.4898          | 0.8      |
| 0.0           | 24.0  | 5160  | 1.5327          | 0.8      |
| 0.0           | 25.0  | 5375  | 1.5448          | 0.8      |
| 0.0           | 26.0  | 5590  | 1.5564          | 0.8      |
| 0.0           | 27.0  | 5805  | 1.5873          | 0.8      |
| 0.0           | 28.0  | 6020  | 1.5904          | 0.8      |
| 0.0           | 29.0  | 6235  | 1.6335          | 0.7778   |
| 0.0           | 30.0  | 6450  | 1.6414          | 0.8      |
| 0.0           | 31.0  | 6665  | 1.6621          | 0.8      |
| 0.0           | 32.0  | 6880  | 1.6886          | 0.7778   |
| 0.0           | 33.0  | 7095  | 1.7082          | 0.7778   |
| 0.0           | 34.0  | 7310  | 1.7346          | 0.7778   |
| 0.0           | 35.0  | 7525  | 1.7374          | 0.7778   |
| 0.0           | 36.0  | 7740  | 1.7722          | 0.7778   |
| 0.0           | 37.0  | 7955  | 1.7793          | 0.7778   |
| 0.0           | 38.0  | 8170  | 1.8303          | 0.7778   |
| 0.0           | 39.0  | 8385  | 1.8111          | 0.7778   |
| 0.0           | 40.0  | 8600  | 1.8625          | 0.7778   |
| 0.0           | 41.0  | 8815  | 1.8394          | 0.7778   |
| 0.0           | 42.0  | 9030  | 1.9233          | 0.7778   |
| 0.0           | 43.0  | 9245  | 1.8877          | 0.7778   |
| 0.0           | 44.0  | 9460  | 1.8931          | 0.7778   |
| 0.0           | 45.0  | 9675  | 1.9380          | 0.7778   |
| 0.0           | 46.0  | 9890  | 1.9422          | 0.7778   |
| 0.0           | 47.0  | 10105 | 1.9560          | 0.7778   |
| 0.0           | 48.0  | 10320 | 1.9635          | 0.7778   |
| 0.0           | 49.0  | 10535 | 1.9545          | 0.7778   |
| 0.0           | 50.0  | 10750 | 1.9649          | 0.7778   |


### Framework versions

- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2