improve readme
Browse files
README.md
CHANGED
@@ -86,36 +86,75 @@ Apart from `img_bytes` these configurations add a total of 18 * 2 additional `in
|
|
86 |
## Dataset Structure
|
87 |
### Data Instances
|
88 |
A sample from the training set is provided below:
|
|
|
|
|
|
|
|
|
89 |
```
|
|
|
|
|
|
|
90 |
{
|
91 |
'img_bytes': b'0x...',
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
92 |
}
|
93 |
```
|
|
|
94 |
### Data Fields
|
95 |
- `img_bytes`: A byte string containing the raw data of a 500x500 PNG image.
|
96 |
|
97 |
-
If `+features` is appended to the dataset name, the following additional fields are
|
98 |
-
|
99 |
-
- `
|
100 |
-
- `
|
101 |
-
- `
|
102 |
-
- `
|
103 |
-
- `
|
104 |
-
- `
|
105 |
-
- `
|
106 |
-
- `
|
107 |
-
- `
|
108 |
-
- `
|
109 |
-
- `
|
110 |
-
- `
|
111 |
-
- `
|
112 |
-
- `
|
113 |
-
- `
|
114 |
-
- `
|
115 |
-
- `
|
116 |
-
- `
|
117 |
-
- `
|
118 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
119 |
### Data Splits
|
120 |
Train
|
121 |
## Dataset Creation
|
|
|
86 |
## Dataset Structure
|
87 |
### Data Instances
|
88 |
A sample from the training set is provided below:
|
89 |
+
```python
|
90 |
+
{
|
91 |
+
'img_bytes': b'0x...',
|
92 |
+
}
|
93 |
```
|
94 |
+
If `+features` is added to the dataset name, the following additional fields are provided:
|
95 |
+
|
96 |
+
```python
|
97 |
{
|
98 |
'img_bytes': b'0x...',
|
99 |
+
'eye_angle': 0,
|
100 |
+
'eye_angle_num_categories': 3,
|
101 |
+
'eye_lashes': 0,
|
102 |
+
'eye_lashes_num_categories': 2,
|
103 |
+
'eye_lid': 0,
|
104 |
+
'eye_lid_num_categories': 2,
|
105 |
+
'chin_length': 2,
|
106 |
+
'chin_length_num_categories': 3,
|
107 |
+
...
|
108 |
}
|
109 |
```
|
110 |
+
|
111 |
### Data Fields
|
112 |
- `img_bytes`: A byte string containing the raw data of a 500x500 PNG image.
|
113 |
|
114 |
+
If `+features` is appended to the dataset name, the following additional `int32` fields are provided:
|
115 |
+
|
116 |
+
- `eye_angle`
|
117 |
+
- `eye_angle_num_categories`
|
118 |
+
- `eye_lashes`
|
119 |
+
- `eye_lashes_num_categories`
|
120 |
+
- `eye_lid`
|
121 |
+
- `eye_lid_num_categories`
|
122 |
+
- `chin_length`
|
123 |
+
- `chin_length_num_categories`
|
124 |
+
- `eyebrow_weight`
|
125 |
+
- `eyebrow_weight_num_categories`
|
126 |
+
- `eyebrow_shape`
|
127 |
+
- `eyebrow_shape_num_categories`
|
128 |
+
- `eyebrow_thickness`
|
129 |
+
- `eyebrow_thickness_num_categories`
|
130 |
+
- `face_shape`
|
131 |
+
- `face_shape_num_categories`
|
132 |
+
- `facial_hair`
|
133 |
+
- `facial_hair_num_categories`
|
134 |
+
- `facial_hair_num_categories`
|
135 |
+
- `facial_hair_num_categories`
|
136 |
+
- `hair`
|
137 |
+
- `hair_num_categories`
|
138 |
+
- `hair_num_categories`
|
139 |
+
- `hair_num_categories`
|
140 |
+
- `eye_color`
|
141 |
+
- `eye_color_num_categories`
|
142 |
+
- `face_color`
|
143 |
+
- `face_color_num_categories`
|
144 |
+
- `hair_color`
|
145 |
+
- `hair_color_num_categories`
|
146 |
+
- `glasses`
|
147 |
+
- `glasses_num_categories`
|
148 |
+
- `glasses_color`
|
149 |
+
- `glasses_color_num_categories`
|
150 |
+
- `eyes_slant`
|
151 |
+
- `eye_slant_num_categories`
|
152 |
+
- `eyebrow_width`
|
153 |
+
- `eyebrow_width_num_categories`
|
154 |
+
- `eye_eyebrow_distance`
|
155 |
+
- `eye_eyebrow_distance_num_categories`
|
156 |
+
|
157 |
+
|
158 |
### Data Splits
|
159 |
Train
|
160 |
## Dataset Creation
|