360CVGroup/FG-CLIP

• hightopics#1

  Add 'bilingual' and 'multilingual' to repository topics

  原因：

  当前

  clip, cross-modal-retrieval, fine-grained-classification, text-image-retrieval

  复制粘贴的修复

  复制

  clip, cross-modal-retrieval, fine-grained-classification, text-image-retrieval, bilingual, multilingual

  待办完成忽略
• mediumabout#2

  Enhance the repository description to highlight bilingual support

  原因：

  当前

  New generation of CLIP with strong fine grained discrimination capability, ICML2026 and ICML2025

  复制粘贴的修复

  复制

  New generation of CLIP for superior fine-grained discrimination and robust bilingual (Chinese/English) vision-language alignment. Accepted at ICML2026 and ICML2025.

  待办完成忽略
• mediumreadme#3

  Refine the README's opening sentence for stronger positioning

  原因：

  当前

  This repository is the official implementation of FG-CLIP and FG-CLIP 2. As a new generation of text-image cross-modal model, it excels in fine-grained understanding. FG-CLIP 2 supports Chinese and English bilingualism, and in 29 datasets and 8 diverse tasks, the model surpasses strong baseline models including SigLIP 2 and MetaCLIP 2, achieving the current best performance in both language tasks.

  复制粘贴的修复

  复制

  FG-CLIP 2 is the official implementation of our next-generation vision-language alignment model, uniquely engineered for **superior fine-grained discrimination** and **robust bilingual (Chinese/English) support**. It significantly outperforms general CLIP models and other strong baselines in detailed text-image understanding across both languages.

  待办完成忽略

本次扫描解析到的品类 GEO 通道：google/gemini-2.5-flash, deepseek/deepseek-v4-flash

• 品类问题

  What are the best models for fine-grained text-image cross-modal retrieval?

  你：未被推荐
  AI 推荐顺序：

  1. CLIP (Contrastive Language-Image Pre-training)
  2. ALBEF (Align before Fuse)
  3. BLIP (Bootstrapping Language-Image Pre-training)
  4. OFA (One-For-All)
  5. CoCa (Contrastive Captioners)
  6. FLAVA (A Foundational Language And Vision Alignment Model)

  AI 推荐了 6 个替代方案，却始终没点名 360CVGroup/FG-CLIP。这就是要补上的差距。

  查看 AI 完整回答
• 品类问题

  Seeking a vision-language alignment model with strong bilingual support for fine-grained tasks.

  你：未被推荐
  AI 推荐顺序：

  1. mPLUG-Owl2
  2. BLIP-2
  3. X-VLM
  4. OpenCLIP
  5. mBERT
  6. XLM-RoBERTa
  7. Flamingo
  8. ViLT

  AI 推荐了 8 个替代方案，却始终没点名 360CVGroup/FG-CLIP。这就是要补上的差距。

  查看 AI 完整回答

• Metadata completeness

  pass

• README presence

  pass

• Compared to common alternatives in this category, what is the core differentiator of 360CVGroup/FG-CLIP?

  pass
  ✓AI 明确点名了 360CVGroup/FG-CLIP

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• If a team adopts 360CVGroup/FG-CLIP in production, what risks or prerequisites should they evaluate first?

  pass
  ✓AI 明确点名了 360CVGroup/FG-CLIP

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• In one sentence, what problem does the repo 360CVGroup/FG-CLIP solve, and who is the primary audience?

  pass
  ✓AI 明确点名了 360CVGroup/FG-CLIP

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？