Datasets:
Hezar: Upload dataset_config.yaml
Browse files- dataset_config.yaml +25 -0
dataset_config.yaml
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
name: image_captioning
|
| 2 |
+
config_type: dataset
|
| 3 |
+
task: image2text
|
| 4 |
+
path: flickr30k-fa
|
| 5 |
+
tokenizer_path: hezarai/roberta-base-fa
|
| 6 |
+
text_column: label
|
| 7 |
+
test_split_size: 0.2
|
| 8 |
+
image_processor_config:
|
| 9 |
+
name: image_processor
|
| 10 |
+
config_type: preprocessor
|
| 11 |
+
mean:
|
| 12 |
+
- 0.5
|
| 13 |
+
- 0.5
|
| 14 |
+
- 0.5
|
| 15 |
+
std:
|
| 16 |
+
- 0.5
|
| 17 |
+
- 0.5
|
| 18 |
+
- 0.5
|
| 19 |
+
rescale: 0.00392156862745098
|
| 20 |
+
resample: 2
|
| 21 |
+
size:
|
| 22 |
+
- 224
|
| 23 |
+
- 224
|
| 24 |
+
mirror: false
|
| 25 |
+
gray_scale: false
|