File size: 3,055 Bytes
0657c20
 
 
 
561fd3c
 
0657c20
 
 
 
 
 
 
 
 
 
 
 
 
561fd3c
0657c20
561fd3c
 
 
0657c20
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
491bbeb
0657c20
 
 
 
 
 
491bbeb
0657c20
 
 
 
 
491bbeb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0657c20
 
 
 
 
 
 
 
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
---
license: apache-2.0
base_model: microsoft/swinv2-tiny-patch4-window8-256
tags:
- image-classification
- vision
- generated_from_trainer
datasets:
- generator
model-index:
- name: swinv2-tiny-panorama-IQA
  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-tiny-panorama-IQA

This model is a fine-tuned version of [microsoft/swinv2-tiny-patch4-window8-256](https://huggingface.co/microsoft/swinv2-tiny-patch4-window8-256) on the isiqa-2019-hf dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0211
- Srocc: 0.0896
- Lcc: 0.2316

## 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: 10
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50.0

### Training results

| Training Loss | Epoch   | Step | Validation Loss | Srocc   | Lcc    |
|:-------------:|:-------:|:----:|:---------------:|:-------:|:------:|
| No log        | 0.8571  | 3    | 0.1747          | 0.1087  | 0.1976 |
| No log        | 2.0     | 7    | 0.0570          | 0.1042  | 0.1898 |
| 0.1599        | 2.8571  | 10   | 0.0320          | 0.1126  | 0.1686 |
| 0.1599        | 4.0     | 14   | 0.0511          | 0.0265  | 0.1235 |
| 0.1599        | 4.8571  | 17   | 0.0274          | -0.0004 | 0.1029 |
| 0.0602        | 6.0     | 21   | 0.0375          | -0.0406 | 0.0900 |
| 0.0602        | 6.8571  | 24   | 0.0306          | -0.0533 | 0.0830 |
| 0.0602        | 8.0     | 28   | 0.0255          | -0.0726 | 0.0714 |
| 0.029         | 8.8571  | 31   | 0.0247          | -0.0568 | 0.0734 |
| 0.029         | 10.0    | 35   | 0.0293          | -0.0429 | 0.0900 |
| 0.029         | 10.8571 | 38   | 0.0259          | -0.0317 | 0.0982 |
| 0.0199        | 12.0    | 42   | 0.0238          | -0.0073 | 0.1288 |
| 0.0199        | 12.8571 | 45   | 0.0243          | 0.0216  | 0.1594 |
| 0.0199        | 14.0    | 49   | 0.0259          | 0.0454  | 0.1810 |
| 0.0161        | 14.8571 | 52   | 0.0224          | 0.0568  | 0.1954 |
| 0.0161        | 16.0    | 56   | 0.0211          | 0.0896  | 0.2316 |
| 0.0161        | 16.8571 | 59   | 0.0223          | 0.1001  | 0.2544 |
| 0.0132        | 18.0    | 63   | 0.0217          | 0.0981  | 0.2681 |
| 0.0132        | 18.8571 | 66   | 0.0221          | 0.1155  | 0.2746 |
| 0.0103        | 20.0    | 70   | 0.0228          | 0.1230  | 0.2831 |
| 0.0103        | 20.8571 | 73   | 0.0245          | 0.1327  | 0.2944 |


### Framework versions

- Transformers 4.42.3
- Pytorch 2.1.2
- Datasets 2.20.0
- Tokenizers 0.19.1