Update README for loading dataset from caption-free branch
Browse filesUpdated the README file to include instructions on loading the COCO 2017 Colorization Dataset from the caption-free branch of the repository. Added usage examples for loading both the train and validation splits from the caption-free branch. The README now provides clear guidance on how to choose between the main branch for original captions and the caption-free branch for prompts-free captions when loading the dataset.
README.md
CHANGED
@@ -91,10 +91,30 @@ unzip train2017.zip
|
|
91 |
unzip val2017.zip
|
92 |
```
|
93 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
94 |
### Loading the Dataset
|
95 |
|
96 |
You can load this dataset using the Hugging Face `datasets` library:
|
97 |
|
|
|
98 |
```python
|
99 |
from datasets import load_dataset
|
100 |
|
@@ -105,6 +125,19 @@ train_dataset = load_dataset("nickpai/coco2017-colorization", split="train")
|
|
105 |
val_dataset = load_dataset("nickpai/coco2017-colorization", split="validation")
|
106 |
```
|
107 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
108 |
## Filtering Criteria
|
109 |
|
110 |
### 1. Grayscale Images
|
|
|
91 |
unzip val2017.zip
|
92 |
```
|
93 |
|
94 |
+
### Branches
|
95 |
+
|
96 |
+
- **main:** Provides the original captions sentences.
|
97 |
+
- **caption-free:** Provides captions with random prompts selected from the following list:
|
98 |
+
|
99 |
+
```python
|
100 |
+
sentences = [
|
101 |
+
"Add colors to this image",
|
102 |
+
"Give realistic colors to this image",
|
103 |
+
"Add realistic colors to this image",
|
104 |
+
"Colorize this grayscale image",
|
105 |
+
"Colorize this image",
|
106 |
+
"Restore the original colors of this image",
|
107 |
+
"Make this image colorful",
|
108 |
+
"Colorize this image as if it was taken with a color camera",
|
109 |
+
"Create the original colors of this image"
|
110 |
+
]
|
111 |
+
```
|
112 |
+
|
113 |
### Loading the Dataset
|
114 |
|
115 |
You can load this dataset using the Hugging Face `datasets` library:
|
116 |
|
117 |
+
#### Main Branch
|
118 |
```python
|
119 |
from datasets import load_dataset
|
120 |
|
|
|
125 |
val_dataset = load_dataset("nickpai/coco2017-colorization", split="validation")
|
126 |
```
|
127 |
|
128 |
+
#### Caption-Free Branch
|
129 |
+
```python
|
130 |
+
from datasets import load_dataset
|
131 |
+
|
132 |
+
# Load the train split of the colorization dataset from the caption-free branch
|
133 |
+
train_dataset = load_dataset("nickpai/coco2017-colorization", split="train", revision="caption-free")
|
134 |
+
|
135 |
+
# Load the validation split of the colorization dataset from the caption-free branch
|
136 |
+
val_dataset = load_dataset("nickpai/coco2017-colorization", split="validation", revision="caption-free")
|
137 |
+
```
|
138 |
+
|
139 |
+
|
140 |
+
|
141 |
## Filtering Criteria
|
142 |
|
143 |
### 1. Grayscale Images
|