{"data":{"slug":"unum-cloud-uform","name":"UForm","tagline":"Pocket-Sized Multimodal AI for content understanding and generation across multilingual texts, images, and video","github_url":"https://github.com/unum-cloud/UForm","owner":"unum-cloud","repo":"UForm","owner_avatar_url":"https://avatars.githubusercontent.com/u/56397513?v=4","primary_language":"Python","stars":1251,"forks":78,"topics":["bert","clip","clustering","contrastive-learning","cross-attention","huggingface-transformers","image-search","language-vision","llava","multi-lingual","multimodal","neural-network","openai","openclip","pretrained-models","pytorch","representation-learning","semantic-search","transformer","vector-search"],"archived":false,"github_pushed_at":"2025-10-30T23:39:54+00:00","maintenance_label":"Slowing","stars_delta_30d":8,"url":"https://www.graphcanon.com/tools/unum-cloud-uform","markdown_url":"https://www.graphcanon.com/tools/unum-cloud-uform.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/unum-cloud-uform","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=unum-cloud-uform","description":"Pocket-Sized Multimodal AI for content understanding and generation across multilingual texts, images, and 🔜 video, up to 5x faster than OpenAI CLIP and LLaVA 🖼️ & 🖋️","homepage_url":"https://unum-cloud.github.io/UForm","license":"Apache-2.0","open_issues":15,"watchers":14,"ai_summary":"A repository for multimodal AI system that focuses on content understanding and generation with capabilities in text, image, and video processing, featuring speed up to 5x faster than OpenAI CLIP and LLaVA.","readme_excerpt":"<h1 align=\"center\">UForm</h1>\n<h3 align=\"center\">\nPocket-Sized Multimodal AI<br/>\nFor Content Understanding and Generation<br/>\n</h3>\n<br/>\n\n<p align=\"center\">\n<a href=\"https://discord.gg/jsMURnSFM2\"><img height=\"25\" src=\"https://github.com/unum-cloud/.github/raw/main/assets/discord.svg\" alt=\"Discord\"></a>\n&nbsp; &nbsp; &nbsp;\n<a href=\"https://www.linkedin.com/company/unum-cloud/\"><img height=\"25\" src=\"https://github.com/unum-cloud/.github/raw/main/assets/linkedin.svg\" alt=\"LinkedIn\"></a>\n&nbsp; &nbsp; &nbsp;\n<a href=\"https://twitter.com/unum_cloud\"><img height=\"25\" src=\"https://github.com/unum-cloud/.github/raw/main/assets/twitter.svg\" alt=\"Twitter\"></a>\n&nbsp; &nbsp; &nbsp;\n<a href=\"https://unum.cloud/post\"><img height=\"25\" src=\"https://github.com/unum-cloud/.github/raw/main/assets/blog.svg\" alt=\"Blog\"></a>\n&nbsp; &nbsp; &nbsp;\n<a href=\"https://github.com/unum-cloud/uform\"><img height=\"25\" src=\"https://github.com/unum-cloud/.github/raw/main/assets/github.svg\" alt=\"GitHub\"></a>\n</p>\n\n<p align=\"center\">\nMultimodal Embeddings from 64 to 768 Dimensions • 1B Parameter Chat\n<br/>\nShort Texts • Images • 🔜 Video Clips • 🔜 Long Documents\n<br/>\nONNX • CoreML • PyTorch\n<br/>\n<a href=\"https://github.com/unum-cloud/uform/blob/main/python/README.md\">Python</a>\n • \n<a href=\"https://github.com/unum-cloud/uform/blob/main/javascript/README.md\">JavaScript</a>\n • \n<a href=\"https://github.com/unum-cloud/uform/blob/main/swift/README.md\">Swift</a>\n</p>\n\n---\n\n\n\nWelcome to UForm, a __multimodal__ AI library that's as versatile as it is efficient.\nUForm [tiny embedding models](#encoder) will help you understand and search visual and textual content across various languages.\nUForm [small generative models](#decoder), on the other hand, don't only support conversational and chat use-cases, but are great for fast image captioning and Visual Question Answering (VQA).\nWith compact __custom pre-trained transformer models__, this can run anywhere from your server farm down to your smartphone.\n\n## Features\n\n- __Tiny Embeddings__: 64-dimensional [Matryoshka][matryoshka]-style embeddings for extremely fast [search][usearch].\n- __Throughput__: Thanks to the small size, the inference speed is [2-4x faster](#speed) than competitors.\n- __Portable__: Models come with native ONNX support, making them easy to deploy on any platform.\n- __Quantization Aware__: Down-cast embeddings from `f32` to `i8` without losing much recall.\n- __Multilingual__: Trained on a balanced dataset, the recall is great across over 20 languages.\n\n[usearch]: https://github.com/unum-cloud/usearch\n[matryoshka]: https://arxiv.org/abs/2205.13147\n\n## Models\n\nFor accuracy and speed benchmarks refer to the [evaluation page](https://github.com/unum-cloud/uform/blob/main/BENCHMARKS.md).\n\n### Embedding Models\n\n<table style=\"width:100%; border-collapse:collapse;\">\n    <thead>\n        <tr>\n            <th>Model</th>\n            <th style=\"text-align:right;\">Parameters</th>\n            <th style=\"text-align:right;\">Languages</th>\n            <th style=\"text-align:right;\">Architecture</th>\n        </tr>\n    </thead>\n    <tbody>\n        <tr>\n            <td><code><a href=\"https://huggingface.co/unum-cloud/uform-vl-english-large/\">uform3-image-text-english-large</a></code>  🆕</td>\n            <td style=\"text-align:right;\">365 M</td>\n            <td style=\"text-align:right;\">1</td>\n            <td style=\"text-align:right;\">12 layer BERT, ViT-L/14</td>\n        </tr>\n        <tr>\n            <td><code><a href=\"https://huggingface.co/unum-cloud/uform-vl-english/\">uform3-image-text-english-base</a></code></td>\n            <td style=\"text-align:right;\">143 M</td>\n            <td style=\"text-align:right;\">1</td>\n            <td style=\"text-align:right;\">4 layer BERT, ViT-B/16</td>\n        </tr>\n        <tr>\n            <td><code><a href=\"https://huggingface.co/unum-cloud/uform-vl-english-small/\">uform3-image-text-english-small</a></code>  🆕</td>\n            <td style=\"text-align:right;\">79 M</td>\n            <td","github_created_at":"2023-02-21T10:04:40+00:00","created_at":"2026-07-11T11:34:09.106093+00:00","updated_at":"2026-09-20T06:22:05.845471+00:00","categories":[{"slug":"computer-vision","name":"Computer Vision","url":"https://www.graphcanon.com/categories/computer-vision","markdown_url":"https://www.graphcanon.com/categories/computer-vision.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/computer-vision"},{"slug":"model-training","name":"Model Training","url":"https://www.graphcanon.com/categories/model-training","markdown_url":"https://www.graphcanon.com/categories/model-training.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/model-training"}],"tags":[{"slug":"bert","name":"bert"},{"slug":"clip","name":"clip"},{"slug":"clustering","name":"clustering"},{"slug":"contrastive-learning","name":"contrastive-learning"},{"slug":"cross-attention","name":"cross-attention"},{"slug":"huggingface-transformers","name":"huggingface-transformers"},{"slug":"image-search","name":"image-search"},{"slug":"language-vision","name":"language-vision"}],"trust":{"provenance":{"is_fork":false,"github_id":604556121,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-20T06:22:03.016Z","maintenance":{"label":"Slowing","score":36,"methodology":"github_public_v1","releases_90d":0,"days_since_push":324,"last_release_at":"2025-10-30T23:39:21Z","stars_delta_30d":8,"open_issues_delta_30d":0},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-11T11:34:10.417Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"mcp":{"source":"repo_scan","observed_at":"2026-09-20T06:22:04.071Z","server_manifest":false},"scan":{"source":"repo_scan","observed_at":"2026-09-20T06:22:04.071Z"},"has_cli":{"value":true,"source":"package.json:bin|scripts","observed_at":"2026-09-20T06:22:04.071Z"},"languages":{"value":["python","javascript"],"source":"github.language+package.json+pyproject.toml","observed_at":"2026-09-20T06:22:04.071Z"},"license_spdx":{"value":"Apache-2.0","source":"github.license","observed_at":"2026-09-20T06:22:04.071Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["Need fast content generation across texts, images, and videos within multilingual contexts.","Prioritize speed in text-image cross-modal tasks where UForm outperforms competing systems."],"when_not_to_use":["Seeking extensive customization options beyond what is offered by UForm's presets.","Requiring a larger contextual understanding model than UForm's compact design provides."],"source":"enrich:decision_facts","observed_at":"2026-07-12T10:53:07.996Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"UForm is a compact multimodal AI system designed for content understanding and generation across multiple languages and media types with faster processing speeds compared to OpenAI CLIP and LLaVA."}]}}