Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -30,4 +30,63 @@ task_categories:
|
|
30 |
pretty_name: FLIR IR YOLO expansion
|
31 |
size_categories:
|
32 |
- 1K<n<10K
|
33 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
pretty_name: FLIR IR YOLO expansion
|
31 |
size_categories:
|
32 |
- 1K<n<10K
|
33 |
+
---
|
34 |
+
|
35 |
+
# FLIR IR YOLO expansion
|
36 |
+
|
37 |
+
Images from repository [FLIR_IR_Expansion](https://github.com/sensationTI/FLIR_IR_Expansion)
|
38 |
+
|
39 |
+
This expansion pack is prepared specifically for training a YOU-ONLY-LOOK-ONCE(YOLO) network. All frames are labeled in the YOLO format. If you want to use this expansion pack for other purposes, images are still available for download but requires manual labeling
|
40 |
+
|
41 |
+
## Download all dataset
|
42 |
+
|
43 |
+
If you want to download all dataset you must do
|
44 |
+
|
45 |
+
```
|
46 |
+
from datasets import load_dataset
|
47 |
+
|
48 |
+
dataset = load_dataset("SAE-AAI/FLIR_IR_Expansion")
|
49 |
+
```
|
50 |
+
|
51 |
+
If you only want to download `train` or `validation` split you must do
|
52 |
+
|
53 |
+
```
|
54 |
+
from datasets import load_dataset
|
55 |
+
|
56 |
+
train_dataset = load_dataset("SAE-AAI/FLIR_IR_Expansion", split='train')
|
57 |
+
validation_dataset = load_dataset("SAE-AAI/FLIR_IR_Expansion", split='validation')
|
58 |
+
```
|
59 |
+
|
60 |
+
## Download by stream
|
61 |
+
|
62 |
+
If you want to donwload dataset by stream, you must do
|
63 |
+
|
64 |
+
```
|
65 |
+
from datasets import load_dataset
|
66 |
+
|
67 |
+
iterable_dataset = load_dataset("SAE-AAI/FLIR_IR_Expansion", streaming=True)
|
68 |
+
```
|
69 |
+
|
70 |
+
now you can get every sample by
|
71 |
+
|
72 |
+
```
|
73 |
+
for sample in iterable_dataset['train']:
|
74 |
+
print(sample['bounding_boxes'], sample['classes'])
|
75 |
+
example['sample'].show()
|
76 |
+
break
|
77 |
+
```
|
78 |
+
|
79 |
+
or with
|
80 |
+
|
81 |
+
```
|
82 |
+
sample = next(iter(iterable_dataset['train']))
|
83 |
+
print(sample['bounding_boxes'], sample['classes'])
|
84 |
+
sample['image']
|
85 |
+
```
|
86 |
+
|
87 |
+
If you want to get a batch of samples you ca do
|
88 |
+
|
89 |
+
```
|
90 |
+
BS = 4
|
91 |
+
example_batch = list(iterable_dataset['train'].take(BS))
|
92 |
+
```
|