Update README.md
Browse files
README.md
CHANGED
@@ -11,7 +11,8 @@ tags:
|
|
11 |
|
12 |
**Scope:** Processing of images, training and evaluation of ViT model,
|
13 |
input file/directory processing, class (category) results of top
|
14 |
-
N predictions output, predictions summarizing into a tabular format
|
|
|
15 |
|
16 |
## Model description:
|
17 |
|
@@ -21,29 +22,29 @@ Training set of the model: **8950** images
|
|
21 |
|
22 |
#### Categories:
|
23 |
|
24 |
-
- **DRAW**: 1182 (11.89%) - drawings, maps, paintings
|
25 |
|
26 |
-
- **DRAW_L**: 813 (8.17%) - drawings, maps, paintings inside tabular layout
|
27 |
|
28 |
-
- **LINE_HW**: 596 (5.99%) - handwritten text lines inside tabular layout
|
29 |
|
30 |
-
- **LINE_P**: 603 (6.06%) - printed text lines inside tabular layout
|
31 |
|
32 |
-
- **LINE_T**: 1332 (13.39%) - typed text lines inside tabular layout
|
33 |
|
34 |
- **PHOTO**: 1015 (10.21%) - photos with text
|
35 |
|
36 |
-
- **PHOTO_L**: 782 (7.86%) - photos inside tabular layout
|
37 |
|
38 |
-
- **TEXT**: 853 (8.58%) - mixed types, printed, and handwritten texts
|
39 |
|
40 |
-
- **TEXT_HW**: 732 (7.36%) - handwritten text
|
41 |
|
42 |
-
- **TEXT_P**: 691 (6.95%) - printed text
|
43 |
|
44 |
-
- **TEXT_T**: 1346 (13.53%) - typed text
|
45 |
|
46 |
-
Evaluation set (10% of the above stats): **995** images
|
47 |
|
48 |
### Result tables:
|
49 |
|
@@ -68,7 +69,7 @@ Page images classification in to 11 predefined categories.
|
|
68 |
|
69 |
#### Preprocessing
|
70 |
|
71 |
-
train_transforms
|
72 |
|
73 |
transforms.RandomApply([
|
74 |
|
@@ -86,37 +87,33 @@ train_transforms = transforms.Compose([
|
|
86 |
|
87 |
], p=0.5),
|
88 |
|
89 |
-
...
|
90 |
|
91 |
-
])
|
92 |
|
93 |
eval_transforms - basic.
|
94 |
|
95 |
#### Training Hyperparameters
|
96 |
|
97 |
-
training_args = TrainingArguments(
|
98 |
|
99 |
-
|
100 |
-
|
101 |
-
|
102 |
-
|
103 |
-
|
104 |
-
|
105 |
-
|
106 |
-
|
107 |
-
|
108 |
-
|
109 |
-
|
110 |
-
|
111 |
-
|
112 |
-
|
113 |
-
|
114 |
-
|
115 |
-
|
116 |
-
|
117 |
-
|
118 |
-
|
119 |
-
)
|
120 |
|
121 |
## Evaluation
|
122 |
|
|
|
11 |
|
12 |
**Scope:** Processing of images, training and evaluation of ViT model,
|
13 |
input file/directory processing, class (category) results of top
|
14 |
+
N predictions output, predictions summarizing into a tabular format,
|
15 |
+
HF hub support for the model
|
16 |
|
17 |
## Model description:
|
18 |
|
|
|
22 |
|
23 |
#### Categories:
|
24 |
|
25 |
+
- **DRAW**: 1182 (11.89%) - drawings, maps, paintings with text
|
26 |
|
27 |
+
- **DRAW_L**: 813 (8.17%) - drawings, maps, paintings with a table legend or inside tabular layout / forms
|
28 |
|
29 |
+
- **LINE_HW**: 596 (5.99%) - handwritten text lines inside tabular layout / forms
|
30 |
|
31 |
+
- **LINE_P**: 603 (6.06%) - printed text lines inside tabular layout / forms
|
32 |
|
33 |
+
- **LINE_T**: 1332 (13.39%) - machine typed text lines inside tabular layout / forms
|
34 |
|
35 |
- **PHOTO**: 1015 (10.21%) - photos with text
|
36 |
|
37 |
+
- **PHOTO_L**: 782 (7.86%) - photos inside tabular layout / forms
|
38 |
|
39 |
+
- **TEXT**: 853 (8.58%) - mixed types, printed, and handwritten texts
|
40 |
|
41 |
+
- **TEXT_HW**: 732 (7.36%) - only handwritten text
|
42 |
|
43 |
+
- **TEXT_P**: 691 (6.95%) - only printed text
|
44 |
|
45 |
+
- **TEXT_T**: 1346 (13.53%) - only machine typed text
|
46 |
|
47 |
+
Evaluation set (10% of the above stats): **995** images
|
48 |
|
49 |
### Result tables:
|
50 |
|
|
|
69 |
|
70 |
#### Preprocessing
|
71 |
|
72 |
+
train_transforms:
|
73 |
|
74 |
transforms.RandomApply([
|
75 |
|
|
|
87 |
|
88 |
], p=0.5),
|
89 |
|
|
|
90 |
|
|
|
91 |
|
92 |
eval_transforms - basic.
|
93 |
|
94 |
#### Training Hyperparameters
|
95 |
|
|
|
96 |
|
97 |
+
eval_strategy="epoch",
|
98 |
+
|
99 |
+
save_strategy="epoch",
|
100 |
+
|
101 |
+
learning_rate=5e-5,
|
102 |
+
|
103 |
+
per_device_train_batch_size=8,
|
104 |
+
|
105 |
+
per_device_eval_batch_size=8,
|
106 |
+
|
107 |
+
num_train_epochs=3,
|
108 |
+
|
109 |
+
warmup_ratio=0.1,
|
110 |
+
|
111 |
+
logging_steps=10,
|
112 |
+
|
113 |
+
load_best_model_at_end=True,
|
114 |
+
|
115 |
+
metric_for_best_model="accuracy",
|
116 |
+
|
|
|
117 |
|
118 |
## Evaluation
|
119 |
|