Document Unsplash media retrieval TODO
Browse files- README.md +10 -3
- README_CN.md +8 -3
README.md
CHANGED
|
@@ -299,11 +299,18 @@ mappings, and metadataset configuration, but not their image files.
|
|
| 299 |
| `partimagenet` | [ImageNet](https://www.image-net.org/download.php) | `train/<wnid>/<wnid>_<id>.JPEG` |
|
| 300 |
| `object365` | [Objects365](https://www.objects365.org/download.html) | `images/train/patch*/objects365_v1_<id>.jpg` |
|
| 301 |
| `sku110k` | [SKU-110K](https://github.com/eg4000/SKU110K_CVPR19) | `images/train_<id>.jpg` |
|
| 302 |
-
| `unsplash` |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 303 |
|
| 304 |
`metadataset.yaml` reads the datasets whose images are already hosted in this
|
| 305 |
-
repository. Use `metadataset-full.yaml` only after all
|
| 306 |
-
|
|
|
|
| 307 |
|
| 308 |
### Prepare an upstream media dataset
|
| 309 |
|
|
|
|
| 299 |
| `partimagenet` | [ImageNet](https://www.image-net.org/download.php) | `train/<wnid>/<wnid>_<id>.JPEG` |
|
| 300 |
| `object365` | [Objects365](https://www.objects365.org/download.html) | `images/train/patch*/objects365_v1_<id>.jpg` |
|
| 301 |
| `sku110k` | [SKU-110K](https://github.com/eg4000/SKU110K_CVPR19) | `images/train_<id>.jpg` |
|
| 302 |
+
| `unsplash` | **TODO — reproducible media retrieval is not yet available** | `raw_unsplash_images/<photo-id>.jpg` |
|
| 303 |
+
|
| 304 |
+
**Unsplash TODO.** The annotations and source-image mappings are included, but
|
| 305 |
+
the current Unsplash distribution does not provide a stable procedure for
|
| 306 |
+
reconstructing the exact historical image bytes referenced by this release.
|
| 307 |
+
Unsplash media hydration is therefore not currently supported. We will document
|
| 308 |
+
a reproducible retrieval and verification procedure when one is available.
|
| 309 |
|
| 310 |
`metadataset.yaml` reads the datasets whose images are already hosted in this
|
| 311 |
+
repository. Use `metadataset-full.yaml` only after all referenced external media
|
| 312 |
+
pools are available locally. Unsplash is excluded from the currently supported
|
| 313 |
+
reproducible hydration workflow.
|
| 314 |
|
| 315 |
### Prepare an upstream media dataset
|
| 316 |
|
README_CN.md
CHANGED
|
@@ -287,10 +287,15 @@ source-image mapping 和 metadataset 配置,但不包含图片文件。
|
|
| 287 |
| `partimagenet` | [ImageNet](https://www.image-net.org/download.php) | `train/<wnid>/<wnid>_<id>.JPEG` |
|
| 288 |
| `object365` | [Objects365](https://www.objects365.org/download.html) | `images/train/patch*/objects365_v1_<id>.jpg` |
|
| 289 |
| `sku110k` | [SKU-110K](https://github.com/eg4000/SKU110K_CVPR19) | `images/train_<id>.jpg` |
|
| 290 |
-
| `unsplash` |
|
| 291 |
|
| 292 |
-
|
| 293 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 294 |
|
| 295 |
### 准备一个上游媒体数据集
|
| 296 |
|
|
|
|
| 287 |
| `partimagenet` | [ImageNet](https://www.image-net.org/download.php) | `train/<wnid>/<wnid>_<id>.JPEG` |
|
| 288 |
| `object365` | [Objects365](https://www.objects365.org/download.html) | `images/train/patch*/objects365_v1_<id>.jpg` |
|
| 289 |
| `sku110k` | [SKU-110K](https://github.com/eg4000/SKU110K_CVPR19) | `images/train_<id>.jpg` |
|
| 290 |
+
| `unsplash` | **TODO——尚无可复现的媒体获取流程** | `raw_unsplash_images/<photo-id>.jpg` |
|
| 291 |
|
| 292 |
+
**Unsplash TODO。** 本仓库已经包含 Unsplash 标注和 source-image mapping,
|
| 293 |
+
但 Unsplash 当前的发布方式无法稳定重建本版本引用的历史图片字节,因此暂不支持
|
| 294 |
+
Unsplash 媒体恢复。待具备可复现的获取与校验方案后,我们会补充相应流程。
|
| 295 |
+
|
| 296 |
+
`metadataset.yaml` 读取图片已经包含在本仓库中的数据集。只有在本地补齐配置
|
| 297 |
+
引用的外部媒体池后,才使用 `metadataset-full.yaml`。当前支持的可复现媒体恢复
|
| 298 |
+
流程不包含 Unsplash。
|
| 299 |
|
| 300 |
### 准备一个上游媒体数据集
|
| 301 |
|