Datasets:
Update README.md
Browse files# PODS: Personal Object Discrimination Suite
<h3 align="center"><a href="https://personalized-rep.github.io" style="color: #2088FF;">🌐Project page</a>           
<a href="https://example2.com" style="color: #2088FF;">📖Paper</a>           
<a href="#citation" style="color: #2088FF;">GitHub</a><br></h3>
We introduce the PODS (Personal Object Discrimination Suite) dataset, a new benchmark for personalized vision tasks.
<p align="center">
<img src="https://cdn-uploads.huggingface.co/production/uploads/65f9d4100f717eb3e67556df/I6COn1U2CmzZFFs998JsL.jpeg" alt="pods.jpg" />
</p>
## PODS
The PODS dataset is new a benchmark for personalized vision tasks. It includes:
* 100 common household objects from 5 semantic categories
* 4 tasks (classification, retrieval, segmentation, detection)
* 4 test splits with different distribution shifts.
* 12 test images per instance (3 per split).
PODS is [available on HuggingFace](#hf-link-here), or can be directly downloaded [here](#link-here).
Metadata is stored in two files:
* `pods_info.json`:
* `classes`: A list of class names
* `class_to_idx`: Mapping of each class to an integer id
* `class_to_sc`: Mapping of each class to a broad, single-word semantic category
* `class_to_split`: Mapping of each class to the `val` or `test` split.
* `pods_image_annos.json`: Maps every image ID to its class and test split (one of `[train, objects, pose, all]`)
|
@@ -31,3 +31,29 @@ dataset_info:
|
|
| 31 |
download_size: 1168203407
|
| 32 |
dataset_size: 1228220099.0
|
| 33 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
download_size: 1168203407
|
| 32 |
dataset_size: 1228220099.0
|
| 33 |
---
|
| 34 |
+
# PODS: Personal Object Discrimination Suite
|
| 35 |
+
<h3 align="center"><a href="https://personalized-rep.github.io" style="color: #2088FF;">🌐Project page</a>           
|
| 36 |
+
<a href="https://example2.com" style="color: #2088FF;">📖Paper</a>           
|
| 37 |
+
<a href="#citation" style="color: #2088FF;">GitHub</a><br></h3>
|
| 38 |
+
|
| 39 |
+
We introduce the PODS (Personal Object Discrimination Suite) dataset, a new benchmark for personalized vision tasks.
|
| 40 |
+
<p align="center">
|
| 41 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/65f9d4100f717eb3e67556df/I6COn1U2CmzZFFs998JsL.jpeg" alt="pods.jpg" />
|
| 42 |
+
</p>
|
| 43 |
+
|
| 44 |
+
## PODS
|
| 45 |
+
The PODS dataset is new a benchmark for personalized vision tasks. It includes:
|
| 46 |
+
* 100 common household objects from 5 semantic categories
|
| 47 |
+
* 4 tasks (classification, retrieval, segmentation, detection)
|
| 48 |
+
* 4 test splits with different distribution shifts.
|
| 49 |
+
* 12 test images per instance (3 per split).
|
| 50 |
+
|
| 51 |
+
PODS is [available on HuggingFace](#hf-link-here), or can be directly downloaded [here](#link-here).
|
| 52 |
+
|
| 53 |
+
Metadata is stored in two files:
|
| 54 |
+
* `pods_info.json`:
|
| 55 |
+
* `classes`: A list of class names
|
| 56 |
+
* `class_to_idx`: Mapping of each class to an integer id
|
| 57 |
+
* `class_to_sc`: Mapping of each class to a broad, single-word semantic category
|
| 58 |
+
* `class_to_split`: Mapping of each class to the `val` or `test` split.
|
| 59 |
+
* `pods_image_annos.json`: Maps every image ID to its class and test split (one of `[train, objects, pose, all]`)
|