Update README.md

README.md

@@ -20,7 +20,7 @@ base_model:
 
 # DanbooruTagQuery
 
-Lightweight multi-label anime image tagger using a pretrained **DINOv3 ViT backbone** and a **cross-attention tag query head**.
+Lightweight multi-label anime image tagger using **DINOv3 ViT backbone** and a **cross-attention tag query head**, trained on Danbooru images.
 
 - 11,424 tags (general + character + copyright)
 - 448×448 input, ONNX-exported
@@ -89,7 +89,7 @@ Key results:
 **Note: this model tags adult (NSFW) content.** Danbooru contains explicit imagery and its tag vocabulary reflects that. Do not use in contexts where such content is unacceptable.
 
 ## Training data
-- Source: [`trojblue/danbooru2025-metadata`](https://huggingface.co/datasets/trojblue/danbooru2025-metadata)
+- Metadata source for Danbooru posts/images: [`trojblue/danbooru2025-metadata`](https://huggingface.co/datasets/trojblue/danbooru2025-metadata)
 - Filters: deleted/banned/flagged/pending images removed, score > 10, file size 50 KB–20 MB, dimensions ≥ 224×224
 - Tags: `tag_string_general` + `tag_string_character` + `tag_string_copyright` unified; tags with frequency ≥ 100; images with ≥ 2 tags
 - Vocabulary: **11,424 tags**; ~1M training images; last 100K images held out as validation