Update README.md

README.md

@@ -7,4 +7,33 @@ language:
 pretty_name: flickr8k
 size_categories:
 - 1K<n<10K
----
+---
+
+You must download the dataset files manually. You can visit [this](https://github.com/jbrownlee/Datasets/releases/tag/Flickr8k) page or run `download.sh` to get files.
+
+After, you can load dataset by referencing the directory:
+
+```py
+import datasets
+ds = datasets.load_dataset("atasoglu/flickr8k-dataset", data_dir="data")
+print(ds)
+```
+
+```
+DatasetDict({
+    train: Dataset({
+        features: ['image_id', 'image_path', 'captions'],
+        num_rows: 6000
+    })
+    test: Dataset({
+        features: ['image_id', 'image_path', 'captions'],
+        num_rows: 1000
+    })
+    validation: Dataset({
+        features: ['image_id', 'image_path', 'captions'],
+        num_rows: 1000
+    })
+})
+```
+
+I don't own the copyright of the images. Please [visit](https://forms.illinois.edu/sec/1713398) for more.