domains / families / media / clip

clip

separate

OpenAI's model connecting vision and language. Enables zero-shot image classification, image-text matching, and cross-modal retrieval. Trained on 400M image-text pairs. Use for image search, content moderation, or vision-language tasks without fine-tuning. Best for general-purpose image understanding.

When to use

• Zero-shot image classification (no training data needed)
• Image-text similarity/matching
• Semantic image search
• Content moderation (detect NSFW, violence)
• Visual question answering
• Cross-modal retrieval (image→text, text→image)
• **25,300+ GitHub stars**
• Trained on 400M image-text pairs

When not to use

• **Not for fine-grained tasks** - Best for broad categories
• **Requires descriptive text** - Vague labels perform poorly
• **Biased on web data** - May have dataset biases
• **No bounding boxes** - Whole image only
• **Limited spatial understanding** - Position/counting weak

Skill page

Source