Hugging Face's logo

DiTo97
/
swinv2-tiny-panorama-IQA

Image Classification
TensorBoard
Safetensors
swinv2
vision
Generated from Trainer
Model card Files Metrics Community
DiTo97 commited on
Commit
491bbeb
•
1 Parent(s): 9ea8096

Model save

Browse files
Files changed (2) hide show
  1. README.md +29 -37
  2. model.safetensors +1 -1
README.md CHANGED
@@ -2,8 +2,6 @@
2
  license: apache-2.0
3
  base_model: microsoft/swinv2-tiny-patch4-window8-256
4
  tags:
5
- - image-classification
6
- - vision
7
  - generated_from_trainer
8
  datasets:
9
  - generator
@@ -17,11 +15,11 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  # swinv2-tiny-panorama-IQA
19
 
20
- 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.
21
  It achieves the following results on the evaluation set:
22
- - Loss: 0.0175
23
- - Srocc: 0.2288
24
- - Lcc: 0.4304
25
 
26
  ## Model description
27
 
@@ -40,48 +38,42 @@ More information needed
40
  ### Training hyperparameters
41
 
42
  The following hyperparameters were used during training:
43
- - learning_rate: 2e-05
44
  - train_batch_size: 32
45
  - eval_batch_size: 32
46
  - seed: 10
47
  - gradient_accumulation_steps: 2
48
  - total_train_batch_size: 64
49
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
- - lr_scheduler_type: linear
51
  - lr_scheduler_warmup_ratio: 0.1
52
  - num_epochs: 50.0
53
 
54
  ### Training results
55
 
56
- | Training Loss | Epoch | Step | Validation Loss | Srocc | Lcc |
57
- |:-------------:|:-------:|:----:|:---------------:|:-------:|:-------:|
58
- | No log | 0.8571 | 3 | 0.1484 | 0.1044 | 0.1970 |
59
- | No log | 2.0 | 7 | 0.0292 | 0.1107 | 0.1713 |
60
- | 0.1442 | 2.8571 | 10 | 0.0548 | 0.0279 | 0.1178 |
61
- | 0.1442 | 4.0 | 14 | 0.0310 | -0.0628 | 0.0689 |
62
- | 0.1442 | 4.8571 | 17 | 0.0476 | -0.0618 | 0.0550 |
63
- | 0.0425 | 6.0 | 21 | 0.0297 | -0.0978 | 0.0051 |
64
- | 0.0425 | 6.8571 | 24 | 0.0259 | -0.0906 | -0.0039 |
65
- | 0.0425 | 8.0 | 28 | 0.0294 | -0.0270 | 0.0079 |
66
- | 0.0242 | 8.8571 | 31 | 0.0237 | -0.0146 | 0.0323 |
67
- | 0.0242 | 10.0 | 35 | 0.0226 | 0.0152 | 0.0809 |
68
- | 0.0242 | 10.8571 | 38 | 0.0236 | 0.0339 | 0.1158 |
69
- | 0.0146 | 12.0 | 42 | 0.0213 | 0.0562 | 0.1753 |
70
- | 0.0146 | 12.8571 | 45 | 0.0199 | 0.0678 | 0.2241 |
71
- | 0.0146 | 14.0 | 49 | 0.0205 | 0.0981 | 0.2702 |
72
- | 0.0116 | 14.8571 | 52 | 0.0190 | 0.1245 | 0.3000 |
73
- | 0.0116 | 16.0 | 56 | 0.0195 | 0.1595 | 0.3511 |
74
- | 0.0116 | 16.8571 | 59 | 0.0194 | 0.1829 | 0.3804 |
75
- | 0.0096 | 18.0 | 63 | 0.0183 | 0.2144 | 0.4030 |
76
- | 0.0096 | 18.8571 | 66 | 0.0195 | 0.2120 | 0.4086 |
77
- | 0.0075 | 20.0 | 70 | 0.0188 | 0.2164 | 0.4127 |
78
- | 0.0075 | 20.8571 | 73 | 0.0209 | 0.2222 | 0.4224 |
79
- | 0.0075 | 22.0 | 77 | 0.0175 | 0.2288 | 0.4304 |
80
- | 0.0076 | 22.8571 | 80 | 0.0211 | 0.2432 | 0.4326 |
81
- | 0.0076 | 24.0 | 84 | 0.0189 | 0.2346 | 0.4327 |
82
- | 0.0076 | 24.8571 | 87 | 0.0188 | 0.2294 | 0.4313 |
83
- | 0.006 | 26.0 | 91 | 0.0223 | 0.2390 | 0.4343 |
84
- | 0.006 | 26.8571 | 94 | 0.0202 | 0.2511 | 0.4399 |
85
 
86
 
87
  ### Framework versions
 
2
  license: apache-2.0
3
  base_model: microsoft/swinv2-tiny-patch4-window8-256
4
  tags:
 
 
5
  - generated_from_trainer
6
  datasets:
7
  - generator
 
15
 
16
  # swinv2-tiny-panorama-IQA
17
 
18
+ 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 generator dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 0.0245
21
+ - Srocc: 0.1327
22
+ - Lcc: 0.2944
23
 
24
  ## Model description
25
 
 
38
  ### Training hyperparameters
39
 
40
  The following hyperparameters were used during training:
41
+ - learning_rate: 1e-05
42
  - train_batch_size: 32
43
  - eval_batch_size: 32
44
  - seed: 10
45
  - gradient_accumulation_steps: 2
46
  - total_train_batch_size: 64
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
+ - lr_scheduler_type: cosine
49
  - lr_scheduler_warmup_ratio: 0.1
50
  - num_epochs: 50.0
51
 
52
  ### Training results
53
 
54
+ | Training Loss | Epoch | Step | Validation Loss | Srocc | Lcc |
55
+ |:-------------:|:-------:|:----:|:---------------:|:-------:|:------:|
56
+ | No log | 0.8571 | 3 | 0.1747 | 0.1087 | 0.1976 |
57
+ | No log | 2.0 | 7 | 0.0570 | 0.1042 | 0.1898 |
58
+ | 0.1599 | 2.8571 | 10 | 0.0320 | 0.1126 | 0.1686 |
59
+ | 0.1599 | 4.0 | 14 | 0.0511 | 0.0265 | 0.1235 |
60
+ | 0.1599 | 4.8571 | 17 | 0.0274 | -0.0004 | 0.1029 |
61
+ | 0.0602 | 6.0 | 21 | 0.0375 | -0.0406 | 0.0900 |
62
+ | 0.0602 | 6.8571 | 24 | 0.0306 | -0.0533 | 0.0830 |
63
+ | 0.0602 | 8.0 | 28 | 0.0255 | -0.0726 | 0.0714 |
64
+ | 0.029 | 8.8571 | 31 | 0.0247 | -0.0568 | 0.0734 |
65
+ | 0.029 | 10.0 | 35 | 0.0293 | -0.0429 | 0.0900 |
66
+ | 0.029 | 10.8571 | 38 | 0.0259 | -0.0317 | 0.0982 |
67
+ | 0.0199 | 12.0 | 42 | 0.0238 | -0.0073 | 0.1288 |
68
+ | 0.0199 | 12.8571 | 45 | 0.0243 | 0.0216 | 0.1594 |
69
+ | 0.0199 | 14.0 | 49 | 0.0259 | 0.0454 | 0.1810 |
70
+ | 0.0161 | 14.8571 | 52 | 0.0224 | 0.0568 | 0.1954 |
71
+ | 0.0161 | 16.0 | 56 | 0.0211 | 0.0896 | 0.2316 |
72
+ | 0.0161 | 16.8571 | 59 | 0.0223 | 0.1001 | 0.2544 |
73
+ | 0.0132 | 18.0 | 63 | 0.0217 | 0.0981 | 0.2681 |
74
+ | 0.0132 | 18.8571 | 66 | 0.0221 | 0.1155 | 0.2746 |
75
+ | 0.0103 | 20.0 | 70 | 0.0228 | 0.1230 | 0.2831 |
76
+ | 0.0103 | 20.8571 | 73 | 0.0245 | 0.1327 | 0.2944 |
 
 
 
 
 
 
77
 
78
 
79
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:531568a6e34a3992dc34d775cfd3c5aa67a8729ab279c6547ac7e0aa90ca75f3
3
  size 110347052
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b175ab7710a8412563774d47532f84babf62db9dfa79d96adde51f802509352a
3
  size 110347052