Classify images and retrieve cross-modal matches using pretrained CLIP models.
npx skills add https://github.com/AlexKoncept/omnia-hub --skill clip-alexkoncept
Please help me install this Agent Skill. Skill: clip Source: https://github.com/AlexKoncept/omnia-hub/tree/main/HERMES/optional-skills/mlops/clip Command: npx skills add https://github.com/AlexKoncept/omnia-hub --skill clip-alexkoncept
CLIP enables machines to understand images and text through a shared embedding space, enabling zero-shot classification and cross-modal retrieval without task-specific fine-tuning.
Run a CLIP workflow to classify or retrieve images by text queries using the preinstalled model.
High-intent search queries and answers about installing and using this skill.
Cross-modal retrieval maps images and text into a shared embedding space to measure similarity directly. This Skill exposes a workflow to embed text queries and image references using pretrained CLIP models, enabling image search by text or text by image.
Yes, pretrained CLIP models support content moderation by classifying images against safety categories zero-shot. This workflow uses transformers, torch, and PIL to embed visual content and compare it against textual moderation criteria.
Zero-shot classification via pretrained vision-language models allows custom categorization without fine-tuning. This Skill provides a Python workflow using CLIP to embed images and your defined text categories, computing similarity for classification.
No, task-specific training data is not required for visual search or image-text similarity. CLIP operates zero-shot by leveraging a shared embedding space to match text queries directly with image references across open-world datasets.
hhhi21gCombine CLIP vision encoders with language models for image understanding tasks.
NVIDIA-TAOMine nearest-neighbor source images against gap-analysis targets for augmentation.
qcmuuEmbed images and text for zero-shot classification and similarity scoring.
NVIDIA-TAOTrain NVDINOv2 vision transformers via self-supervised learning without labeled data.
