Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -65,13 +65,16 @@ Researchers and developers can utilize this dataset for various tasks such as:
|
|
65 |
- Text length imbalance: the longest text has the length of 8903 whereas the shortest is 1. This can create a situation of highly ram usage with using LSTM model,etc..
|
66 |
|
67 |
### View dataset:
|
68 |
-
|
|
|
|
|
69 |
```python
|
70 |
from datasets import load_dataset
|
71 |
dataset = load_dataset("Seeker38/image_text_wikipedia_vi", split="train")
|
72 |
```
|
|
|
73 |
|
74 |
-
<b>For dataset that has been downloaded to local</b>
|
75 |
```python
|
76 |
import pandas as pd
|
77 |
from datasets import Dataset
|
@@ -82,15 +85,19 @@ df = pd.read_parquet(parquet_file)
|
|
82 |
# Convert the pandas DataFrame to a datasets.arrow_dataset.Dataset object
|
83 |
dataset = Dataset.from_pandas(df)
|
84 |
```
|
85 |
-
|
|
|
86 |
```python
|
|
|
87 |
dataset[3]["text"]
|
88 |
```
|
89 |
-
|
|
|
90 |
```python
|
91 |
from PIL import Image
|
92 |
import io
|
93 |
|
|
|
94 |
image_bytes = dataset[3]["image"]["bytes"]
|
95 |
|
96 |
# Convert bytes to Image
|
@@ -100,4 +107,7 @@ image_rgb = image.convert("RGB") # some images have error: ValueError: Could not
|
|
100 |
image_rgb
|
101 |
```
|
102 |
|
103 |
-
|
|
|
|
|
|
|
|
65 |
- Text length imbalance: the longest text has the length of 8903 whereas the shortest is 1. This can create a situation of highly ram usage with using LSTM model,etc..
|
66 |
|
67 |
### View dataset:
|
68 |
+
There are 2 ways to load dataset:
|
69 |
+
|
70 |
+
<b>1. Use datasets library instead of downloading the dataset to local</b>
|
71 |
```python
|
72 |
from datasets import load_dataset
|
73 |
dataset = load_dataset("Seeker38/image_text_wikipedia_vi", split="train")
|
74 |
```
|
75 |
+
##### you can use the link from this <b>[Google Colab](https://colab.research.google.com/drive/1BOAEsiVXNGm__vhZ4v_oyqytweG3JTm_?usp=sharing)</b> to see a little viewing demo.
|
76 |
|
77 |
+
<b>2. For dataset that has been downloaded to local</b>
|
78 |
```python
|
79 |
import pandas as pd
|
80 |
from datasets import Dataset
|
|
|
85 |
# Convert the pandas DataFrame to a datasets.arrow_dataset.Dataset object
|
86 |
dataset = Dataset.from_pandas(df)
|
87 |
```
|
88 |
+
|
89 |
+
<b>To view the element's text</b>
|
90 |
```python
|
91 |
+
# Example: element number 3
|
92 |
dataset[3]["text"]
|
93 |
```
|
94 |
+
|
95 |
+
<b>If you use the 2nd way, then to view,or even use for training the element's image, you need to contain the convertion step</b>
|
96 |
```python
|
97 |
from PIL import Image
|
98 |
import io
|
99 |
|
100 |
+
# Example: element number 3
|
101 |
image_bytes = dataset[3]["image"]["bytes"]
|
102 |
|
103 |
# Convert bytes to Image
|
|
|
107 |
image_rgb
|
108 |
```
|
109 |
|
110 |
+
<b>Else</b>
|
111 |
+
```python
|
112 |
+
dataset[2]["image"]
|
113 |
+
```
|