File size: 5,553 Bytes
46d6868
 
 
 
9b680ae
46d6868
 
 
 
 
 
 
 
 
 
 
9b680ae
46d6868
9b680ae
 
 
 
46d6868
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
base_model: google/vit-base-patch16-224-in21k
tags:
- image-classification
- generated_from_trainer
model-index:
- name: ryan03312024_lr_2e-5_wd_001
  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. -->

# ryan03312024_lr_2e-5_wd_001

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 properties dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1916
- Ordinal Mae: 0.4221
- Ordinal Accuracy: 0.6828
- Na Accuracy: 0.8591

## 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: 2e-05
- 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: 1.5
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Ordinal Mae | Ordinal Accuracy | Na Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:-----------:|:----------------:|:-----------:|
| 0.4436        | 0.04  | 100  | 0.3698          | 0.8706      | 0.3332           | 0.7990      |
| 0.3143        | 0.07  | 200  | 0.3215          | 0.8555      | 0.4017           | 0.8093      |
| 0.3385        | 0.11  | 300  | 0.2997          | 0.8303      | 0.4485           | 0.8591      |
| 0.3127        | 0.14  | 400  | 0.2889          | 0.8013      | 0.4881           | 0.8746      |
| 0.3054        | 0.18  | 500  | 0.2804          | 0.7619      | 0.5325           | 0.8780      |
| 0.3051        | 0.22  | 600  | 0.2752          | 0.7215      | 0.5235           | 0.9158      |
| 0.2833        | 0.25  | 700  | 0.2653          | 0.6807      | 0.5487           | 0.8969      |
| 0.2907        | 0.29  | 800  | 0.2550          | 0.6432      | 0.5618           | 0.8351      |
| 0.2468        | 0.32  | 900  | 0.2522          | 0.6119      | 0.5972           | 0.8058      |
| 0.2199        | 0.36  | 1000 | 0.2437          | 0.6023      | 0.6062           | 0.8127      |
| 0.2219        | 0.4   | 1100 | 0.2361          | 0.5574      | 0.5959           | 0.9038      |
| 0.2071        | 0.43  | 1200 | 0.2387          | 0.5439      | 0.6175           | 0.7715      |
| 0.2214        | 0.47  | 1300 | 0.2341          | 0.5257      | 0.6232           | 0.7955      |
| 0.2627        | 0.5   | 1400 | 0.2315          | 0.5152      | 0.6124           | 0.7990      |
| 0.2067        | 0.54  | 1500 | 0.2247          | 0.5026      | 0.6396           | 0.8110      |
| 0.2086        | 0.58  | 1600 | 0.2192          | 0.4955      | 0.6589           | 0.8041      |
| 0.1993        | 0.61  | 1700 | 0.2182          | 0.4738      | 0.6522           | 0.8127      |
| 0.1962        | 0.65  | 1800 | 0.2211          | 0.4858      | 0.6232           | 0.9141      |
| 0.1882        | 0.69  | 1900 | 0.2045          | 0.4669      | 0.6632           | 0.8625      |
| 0.1895        | 0.72  | 2000 | 0.2082          | 0.4696      | 0.6316           | 0.8608      |
| 0.1979        | 0.76  | 2100 | 0.2270          | 0.4791      | 0.6373           | 0.9003      |
| 0.2643        | 0.79  | 2200 | 0.2069          | 0.4663      | 0.6414           | 0.8557      |
| 0.2279        | 0.83  | 2300 | 0.2030          | 0.4581      | 0.6543           | 0.8694      |
| 0.1965        | 0.87  | 2400 | 0.2109          | 0.4446      | 0.6820           | 0.8007      |
| 0.1637        | 0.9   | 2500 | 0.2005          | 0.4439      | 0.6763           | 0.8557      |
| 0.1705        | 0.94  | 2600 | 0.1964          | 0.4321      | 0.6748           | 0.8540      |
| 0.2412        | 0.97  | 2700 | 0.1958          | 0.4345      | 0.6730           | 0.8780      |
| 0.1438        | 1.01  | 2800 | 0.1972          | 0.4301      | 0.6784           | 0.8471      |
| 0.123         | 1.05  | 2900 | 0.1995          | 0.4231      | 0.6753           | 0.8419      |
| 0.1411        | 1.08  | 3000 | 0.1946          | 0.4220      | 0.6817           | 0.8454      |
| 0.1443        | 1.12  | 3100 | 0.1916          | 0.4221      | 0.6828           | 0.8591      |
| 0.208         | 1.15  | 3200 | 0.1942          | 0.4163      | 0.6740           | 0.8677      |
| 0.1343        | 1.19  | 3300 | 0.1962          | 0.4182      | 0.6889           | 0.8471      |
| 0.1347        | 1.23  | 3400 | 0.1938          | 0.4161      | 0.6900           | 0.8660      |
| 0.1076        | 1.26  | 3500 | 0.1970          | 0.4181      | 0.6943           | 0.8471      |
| 0.1248        | 1.3   | 3600 | 0.1951          | 0.4151      | 0.6959           | 0.8471      |
| 0.1455        | 1.33  | 3700 | 0.1952          | 0.4147      | 0.6851           | 0.8814      |
| 0.131         | 1.37  | 3800 | 0.1953          | 0.4172      | 0.6948           | 0.8454      |
| 0.1307        | 1.41  | 3900 | 0.1932          | 0.4127      | 0.6928           | 0.8643      |
| 0.1198        | 1.44  | 4000 | 0.1947          | 0.4110      | 0.6941           | 0.8574      |
| 0.1363        | 1.48  | 4100 | 0.1952          | 0.4087      | 0.6887           | 0.8574      |


### Framework versions

- Transformers 4.39.1
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2