File size: 4,513 Bytes
ca5cfe6
 
 
 
e141d46
 
ca5cfe6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e141d46
ca5cfe6
e141d46
 
 
 
 
ca5cfe6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
base_model: microsoft/swinv2-large-patch4-window12-192-22k
tags:
- image-classification
- vision
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: swinv2-large-patch4-window12-192-22k-finetuned-galaxy10-decals
  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. -->

# swinv2-large-patch4-window12-192-22k-finetuned-galaxy10-decals

This model is a fine-tuned version of [microsoft/swinv2-large-patch4-window12-192-22k](https://huggingface.co/microsoft/swinv2-large-patch4-window12-192-22k) on the matthieulel/galaxy10_decals dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4372
- Accuracy: 0.8568
- Precision: 0.8575
- Recall: 0.8568
- F1: 0.8550

## 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: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 30

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1     |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| 0.974         | 0.99  | 62   | 0.7350          | 0.7480   | 0.7464    | 0.7480 | 0.7365 |
| 0.7716        | 2.0   | 125  | 0.6093          | 0.7982   | 0.8102    | 0.7982 | 0.7960 |
| 0.6813        | 2.99  | 187  | 0.5034          | 0.8286   | 0.8301    | 0.8286 | 0.8254 |
| 0.5998        | 4.0   | 250  | 0.4645          | 0.8433   | 0.8431    | 0.8433 | 0.8403 |
| 0.5306        | 4.99  | 312  | 0.4889          | 0.8320   | 0.8377    | 0.8320 | 0.8336 |
| 0.5234        | 6.0   | 375  | 0.5036          | 0.8309   | 0.8398    | 0.8309 | 0.8278 |
| 0.4984        | 6.99  | 437  | 0.4482          | 0.8478   | 0.8484    | 0.8478 | 0.8461 |
| 0.456         | 8.0   | 500  | 0.4370          | 0.8557   | 0.8573    | 0.8557 | 0.8557 |
| 0.4672        | 8.99  | 562  | 0.4372          | 0.8568   | 0.8575    | 0.8568 | 0.8550 |
| 0.4211        | 10.0  | 625  | 0.4428          | 0.8523   | 0.8513    | 0.8523 | 0.8505 |
| 0.4228        | 10.99 | 687  | 0.4762          | 0.8433   | 0.8459    | 0.8433 | 0.8435 |
| 0.3966        | 12.0  | 750  | 0.4943          | 0.8410   | 0.8434    | 0.8410 | 0.8404 |
| 0.383         | 12.99 | 812  | 0.4885          | 0.8478   | 0.8503    | 0.8478 | 0.8463 |
| 0.3899        | 14.0  | 875  | 0.5021          | 0.8472   | 0.8494    | 0.8472 | 0.8474 |
| 0.3364        | 14.99 | 937  | 0.5107          | 0.8495   | 0.8488    | 0.8495 | 0.8486 |
| 0.331         | 16.0  | 1000 | 0.5219          | 0.8484   | 0.8460    | 0.8484 | 0.8454 |
| 0.288         | 16.99 | 1062 | 0.5696          | 0.8422   | 0.8429    | 0.8422 | 0.8410 |
| 0.2867        | 18.0  | 1125 | 0.5529          | 0.8484   | 0.8474    | 0.8484 | 0.8473 |
| 0.2889        | 18.99 | 1187 | 0.5613          | 0.8529   | 0.8522    | 0.8529 | 0.8520 |
| 0.2809        | 20.0  | 1250 | 0.6093          | 0.8433   | 0.8378    | 0.8433 | 0.8391 |
| 0.2684        | 20.99 | 1312 | 0.6096          | 0.8444   | 0.8409    | 0.8444 | 0.8419 |
| 0.2809        | 22.0  | 1375 | 0.6100          | 0.8455   | 0.8453    | 0.8455 | 0.8445 |
| 0.2661        | 22.99 | 1437 | 0.6161          | 0.8354   | 0.8378    | 0.8354 | 0.8359 |
| 0.2435        | 24.0  | 1500 | 0.6540          | 0.8517   | 0.8512    | 0.8517 | 0.8512 |
| 0.2593        | 24.99 | 1562 | 0.6644          | 0.8472   | 0.8462    | 0.8472 | 0.8456 |
| 0.2343        | 26.0  | 1625 | 0.6655          | 0.8467   | 0.8441    | 0.8467 | 0.8449 |
| 0.2281        | 26.99 | 1687 | 0.6759          | 0.8450   | 0.8438    | 0.8450 | 0.8440 |
| 0.2334        | 28.0  | 1750 | 0.6836          | 0.8472   | 0.8445    | 0.8472 | 0.8451 |
| 0.2129        | 28.99 | 1812 | 0.6731          | 0.8489   | 0.8466    | 0.8489 | 0.8471 |
| 0.2252        | 29.76 | 1860 | 0.6773          | 0.8467   | 0.8440    | 0.8467 | 0.8449 |


### Framework versions

- Transformers 4.37.2
- Pytorch 2.3.0
- Datasets 2.19.1
- Tokenizers 0.15.1