[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6AE1D0YV800DDPVAPEKM0C6":3,"skill-same:01M6AE1D0YV800DDPVAPEKM0C6":71},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":42,"linkOnly":43,"markdownContent":44,"markdownBody":45,"frontMatter":46,"refName":47,"commitSha":48,"skillDirectory":49,"forksCount":50,"repositoryHtmlUrl":51,"repositoryHomepageUrl":52,"ownerHtmlUrl":53,"ownerBlogUrl":54,"ownerDescription":55,"updatedAt":56,"install":57,"resources":70},"01M6AE1D0YV800DDPVAPEKM0C6","qdrant-model-migration",null,"Guides embedding model migration in Qdrant without downtime. Use when someone asks 'how to switch embedding models', 'how to migrate vectors', 'how to update to a new model', 'zero-downtime model change', 'how to re-embed my data', or 'can I use two models at once'. Also use when upgrading model dimensions, switching providers, or A\u002FB testing models.","github\u002Fawesome-copilot","github","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F9919?v=4","01M6AC83DS133TPXFSP1DHNCC6","GitHub","01M6AD3PBDYPK7NVFJNJ66Y8RS",41046,"MIT","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fqdrant-model-migration\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41],"Qdrant","qdrant model migration","model","migration","switch","new","when","can","re","embedding","embed","guides","how to","switching","zero-downtime","dimensions","testing","use","migrate","data","ZeRO","update","change","2026-11-10 10:12:30",false,"---\nname: qdrant-model-migration\ndescription: \"Guides embedding model migration in Qdrant without downtime. Use when someone asks 'how to switch embedding models', 'how to migrate vectors', 'how to update to a new model', 'zero-downtime model change', 'how to re-embed my data', or 'can I use two models at once'. Also use when upgrading model dimensions, switching providers, or A\u002FB testing models.\"\n---\n\n# What to Do When Changing Embedding Models\n\nVectors from different models are incompatible. You cannot mix old and new embeddings in the same vector space. You also cannot add new named vector fields to an existing collection. All named vectors must be defined at collection creation time. Both migration strategies below require creating a new collection.\n\n- Understand collection aliases before choosing a strategy [Collection aliases](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fcollections\u002F?s=collection-aliases)\n\n\n## Can I Avoid Re-embedding?\n\nUse when: looking for shortcuts before committing to full migration.\n\nYou MUST re-embed if: changing model provider (OpenAI to Cohere), changing architecture (CLIP to BGE), incompatible dimension counts across different models, or adding sparse vectors to dense-only collection.\n\nYou CAN avoid re-embedding if: using Matryoshka models (use `dimensions` parameter to output lower-dimensional embeddings, learn linear transformation from sample data, some recall loss, good for 100M+ datasets). Or changing quantization (binary to scalar): Qdrant re-quantizes automatically. [Quantization](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fquantization\u002F)\n\n\n## Need Zero Downtime (Alias Swap)\n\nUse when: production must stay available. Recommended for model replacement at scale.\n\n- Create a new collection with the new model's dimensions and distance metric\n- Re-embed all data into the new collection in the background\n- Point your application at a collection alias instead of a direct collection name\n- Atomically swap the alias to the new collection [Switch collection](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fcollections\u002F?s=switch-collection)\n- Verify search quality, then delete the old collection\n\nCareful, the alias swap only redirects queries. Payloads must be re-uploaded separately.\n\n\n## Need Both Models Live (Side-by-Side)\n\nUse when: A\u002FB testing models, multi-modal (dense + sparse), or evaluating a new model before committing.\n\nYou cannot add a named vector to an existing collection. Create a new collection with both vector fields defined upfront:\n\n- Create new collection with old and new named vectors both defined [Collection with multiple vectors](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fcollections\u002F?s=collection-with-multiple-vectors)\n- Migrate data from old collection, preserving existing vectors in the old named field\n- Backfill new model embeddings incrementally using `UpdateVectors` [Update vectors](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fpoints\u002F?s=update-vectors)\n- Compare quality by querying with `using: \"old_model\"` vs `using: \"new_model\"`\n- Swap alias to new collection once satisfied\n\nCo-locating large multi-vectors (especially ColBERT) with dense vectors degrades ALL queries, even those only using dense. At millions of points, users report 13s latency dropping to 2s after removing ColBERT. Put large vectors on disk during side-by-side migration.\n\nIf you anticipate future model migrations, define both vector fields upfront at collection creation.\n\n\n## Dense to Hybrid Search Migration\n\nUse when: adding sparse\u002FBM25 vectors to an existing dense-only collection. Most common migration pattern.\n\nYou cannot add sparse vectors to an existing dense-only collection. Must recreate:\n\n- Create new collection with both dense and sparse vector configs defined\n- Re-embed all data with both dense and sparse models\n- Migrate payloads, swap alias\n\nSparse vectors at chunk level have different TF-IDF characteristics than document level. Test retrieval quality after migration, especially for non-English text without stop-word removal.\n\n\n## Re-embedding Is Too Slow\n\nUse when: dataset is large and re-embedding is the bottleneck.\n\n- Use `update_mode: insert` (v1.17+) for safe idempotent migration [Update mode](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fpoints\u002F?s=update-mode)\n- Scroll the old collection with `with_vectors=False`, re-embed in batches, upsert into new collection\n- Upload in parallel batches (64-256 points per request, 2-4 parallel streams) [Bulk upload](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Ftutorials-develop\u002Fbulk-upload\u002F)\n- Disable HNSW during bulk load (set `indexing_threshold_kb` very high, restore after)\n- For Qdrant Cloud inference, switching models is a config change, not a pipeline change [Inference docs](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Finference\u002F)\n\nFor 400GB+ datasets, expect days. For small datasets (\u003C25MB), re-indexing from source is faster than using the migration tool.\n\n\n## What NOT to Do\n\n- Assume you can add named vectors to an existing collection (must be defined at creation time)\n- Delete the old collection before verifying the new one\n- Forget to update the query embedding model in your application code\n- Skip payload migration when using alias swap (aliases redirect queries, they do not copy data)\n- Keep ColBERT vectors co-located with dense vectors during a long migration (I\u002FO cost degrades all queries)\n- Migrate to hybrid search without testing BM25 quality at chunk level\n","# What to Do When Changing Embedding Models\n\nVectors from different models are incompatible. You cannot mix old and new embeddings in the same vector space. You also cannot add new named vector fields to an existing collection. All named vectors must be defined at collection creation time. Both migration strategies below require creating a new collection.\n\n- Understand collection aliases before choosing a strategy [Collection aliases](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fcollections\u002F?s=collection-aliases)\n\n\n## Can I Avoid Re-embedding?\n\nUse when: looking for shortcuts before committing to full migration.\n\nYou MUST re-embed if: changing model provider (OpenAI to Cohere), changing architecture (CLIP to BGE), incompatible dimension counts across different models, or adding sparse vectors to dense-only collection.\n\nYou CAN avoid re-embedding if: using Matryoshka models (use `dimensions` parameter to output lower-dimensional embeddings, learn linear transformation from sample data, some recall loss, good for 100M+ datasets). Or changing quantization (binary to scalar): Qdrant re-quantizes automatically. [Quantization](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fquantization\u002F)\n\n\n## Need Zero Downtime (Alias Swap)\n\nUse when: production must stay available. Recommended for model replacement at scale.\n\n- Create a new collection with the new model's dimensions and distance metric\n- Re-embed all data into the new collection in the background\n- Point your application at a collection alias instead of a direct collection name\n- Atomically swap the alias to the new collection [Switch collection](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fcollections\u002F?s=switch-collection)\n- Verify search quality, then delete the old collection\n\nCareful, the alias swap only redirects queries. Payloads must be re-uploaded separately.\n\n\n## Need Both Models Live (Side-by-Side)\n\nUse when: A\u002FB testing models, multi-modal (dense + sparse), or evaluating a new model before committing.\n\nYou cannot add a named vector to an existing collection. Create a new collection with both vector fields defined upfront:\n\n- Create new collection with old and new named vectors both defined [Collection with multiple vectors](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fcollections\u002F?s=collection-with-multiple-vectors)\n- Migrate data from old collection, preserving existing vectors in the old named field\n- Backfill new model embeddings incrementally using `UpdateVectors` [Update vectors](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fpoints\u002F?s=update-vectors)\n- Compare quality by querying with `using: \"old_model\"` vs `using: \"new_model\"`\n- Swap alias to new collection once satisfied\n\nCo-locating large multi-vectors (especially ColBERT) with dense vectors degrades ALL queries, even those only using dense. At millions of points, users report 13s latency dropping to 2s after removing ColBERT. Put large vectors on disk during side-by-side migration.\n\nIf you anticipate future model migrations, define both vector fields upfront at collection creation.\n\n\n## Dense to Hybrid Search Migration\n\nUse when: adding sparse\u002FBM25 vectors to an existing dense-only collection. Most common migration pattern.\n\nYou cannot add sparse vectors to an existing dense-only collection. Must recreate:\n\n- Create new collection with both dense and sparse vector configs defined\n- Re-embed all data with both dense and sparse models\n- Migrate payloads, swap alias\n\nSparse vectors at chunk level have different TF-IDF characteristics than document level. Test retrieval quality after migration, especially for non-English text without stop-word removal.\n\n\n## Re-embedding Is Too Slow\n\nUse when: dataset is large and re-embedding is the bottleneck.\n\n- Use `update_mode: insert` (v1.17+) for safe idempotent migration [Update mode](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fpoints\u002F?s=update-mode)\n- Scroll the old collection with `with_vectors=False`, re-embed in batches, upsert into new collection\n- Upload in parallel batches (64-256 points per request, 2-4 parallel streams) [Bulk upload](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Ftutorials-develop\u002Fbulk-upload\u002F)\n- Disable HNSW during bulk load (set `indexing_threshold_kb` very high, restore after)\n- For Qdrant Cloud inference, switching models is a config change, not a pipeline change [Inference docs](https:\u002F\u002Fsearch.qdrant.tech\u002Fmd\u002Fdocumentation\u002Finference\u002F)\n\nFor 400GB+ datasets, expect days. For small datasets (\u003C25MB), re-indexing from source is faster than using the migration tool.\n\n\n## What NOT to Do\n\n- Assume you can add named vectors to an existing collection (must be defined at creation time)\n- Delete the old collection before verifying the new one\n- Forget to update the query embedding model in your application code\n- Skip payload migration when using alias swap (aliases redirect queries, they do not copy data)\n- Keep ColBERT vectors co-located with dense vectors during a long migration (I\u002FO cost degrades all queries)\n- Migrate to hybrid search without testing BM25 quality at chunk level\n",{"name":5,"description":7},"main","5269f7ea2d6c3a9bfda5ed878b6726cf509121fd","skills\u002Fqdrant-model-migration",5340,"https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot","https:\u002F\u002Fawesome-copilot.github.com\u002F","https:\u002F\u002Fgithub.com\u002Fgithub","https:\u002F\u002Fgithub.com\u002Fabout","How people build software.","2026-11-11 13:41:32",{"source":8,"skillName":5,"commands":58,"prompt":68,"zipAvailable":69},[59,62,65],{"tool":60,"command":61},"npx","npx skills add github\u002Fawesome-copilot --skill qdrant-model-migration",{"tool":63,"command":64},"bunx","bunx skills add github\u002Fawesome-copilot --skill qdrant-model-migration",{"tool":66,"command":67},"pnpm","pnpm dlx skills add github\u002Fawesome-copilot --skill qdrant-model-migration","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fqdrant-model-migration\u002FSKILL.md",true,[],[72,104,132,161,195,212],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":86,"publishedAt":42},"01M70KXZNX9ES38N2A8M58ZZMZ","supabase-rls-policies","Write, review, and test Supabase Row Level Security (RLS) policies so that one user cannot read or change another user's rows. Use when creating or changing tables, policies, views, or security definer functions in a Supabase project, when reviewing a migration that touches RLS, or when asked to prove user isolation with pgTAP tests that sign in as two users.","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fsupabase-rls-policies\u002FSKILL.md",[78,83],{"categoryNo":79,"categoryName":80,"taxonomyType":81,"taxonomyTypeDesc":82},"01JBSEEDCAT000000000000008","测试与质量","SCENARIO","使用场景",{"categoryNo":84,"categoryName":85,"taxonomyType":81,"taxonomyTypeDesc":82},"01JBSEEDCAT000000000000023","Skill与智能体开发",[87,88,89,90,91,25,92,41,93,94,95,96,97,98,99,100,36,101,102,103],"supabase","rls","supabase rls","supabase rls policies","project","review","read","write","security","isolation","tables","as","tests","prove","users","functions","with",{"skillNo":105,"name":106,"nameZh":6,"description":107,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":108,"categories":109,"keywords":111,"publishedAt":42},"01M70KWA5YWXDV6V8DZ19E3T2Y","skill-distribution-audit","Audit a skill collection before release or investigate installer omissions, archived\u002Ftemplate leakage, duplicate identities, and manifest drift. Reconcile the intended public source set, distribution manifest, and actual discovery output for a specified consumer version. Includes a deterministic set-comparison helper.","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fskill-distribution-audit\u002FSKILL.md",[110],{"categoryNo":84,"categoryName":85,"taxonomyType":81,"taxonomyTypeDesc":82},[112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131],"distribution","skill distribution audit","audit","SET","deterministic","installer","for","discovery","comparison","helper","manifest","version","collection","drift","output","investigate","duplicate","template","release","reconcile",{"skillNo":133,"name":134,"nameZh":6,"description":135,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":136,"sourceHtmlUrl":137,"categories":138,"keywords":144,"publishedAt":42},"01M70KCJPFVW7PPQ49CC5R8JGK","logicapps-custom-code-net10-migration","Migrate Azure Logic Apps Standard local-function custom-code projects from .NET 8 to .NET 10. Use for legacy Logic Apps custom-code workspace upgrades, project\u002Fpackage configuration, custom-worker settings, publication, and integrated VS Code debugging. Not for ordinary .NET applications, standalone Azure Functions, Consumption workflows, or cloud deployment.","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Flogicapps-custom-code-net10-migration\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Flogicapps-custom-code-net10-migration\u002FSKILL.md",[139],{"categoryNo":140,"categoryName":141,"taxonomyType":142,"taxonomyTypeDesc":143},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[145,146,147,22,148,149,118,91,150,151,152,153,154,155,156,36,157,158,159,160],"custom","code","logicapps custom code net10 migration","projects","standalone","deployment","applications","local","azure","worker","upgrades","package","apps","Legacy","debugging","workflows",{"skillNo":162,"name":163,"nameZh":6,"description":164,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":165,"sourceHtmlUrl":166,"categories":167,"keywords":172,"publishedAt":42},"01M6AEBN37RQMXJ3298PEY7PG1","winmd-api-search","Find and explore Windows desktop APIs. Use when building features that need platform capabilities — camera, file access, notifications, UI controls, AI\u002FML, sensors, networking, etc. Discovers the right API for a task and retrieves full type details (methods, properties, events, enumeration values).","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fwinmd-api-search\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fwinmd-api-search\u002FSKILL.md",[168,171],{"categoryNo":169,"categoryName":170,"taxonomyType":81,"taxonomyTypeDesc":82},"01JBSEEDCAT000000000000010","后端与接口",{"categoryNo":140,"categoryName":141,"taxonomyType":142,"taxonomyTypeDesc":143},[173,174,175,176,177,178,179,180,181,182,183,184,185,186,187,25,188,189,190,191,192,118,193,194],"winmd api search","search","api","windows","desktop","file access","enumeration","ui","ml","file","networking","capabilities","notifications","explore","access","features","full","camera","events","platform","methods","controls",{"skillNo":196,"name":197,"nameZh":6,"description":198,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":199,"licenseFileUrl":200,"sourceHtmlUrl":201,"categories":202,"keywords":208,"publishedAt":42},"01M6AE571R3SC1ZGE202B3NB5S","resemble-detect","Deepfake detection and media safety — detect AI-generated audio, images, video, and text, trace synthesis sources, apply watermarks, verify speaker identity, and analyze media intelligence using Resemble AI","Apache-2.0","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fresemble-detect\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fresemble-detect\u002FSKILL.md",[203,204,207],{"categoryNo":169,"categoryName":170,"taxonomyType":81,"taxonomyTypeDesc":82},{"categoryNo":205,"categoryName":206,"taxonomyType":81,"taxonomyTypeDesc":82},"01JBSEEDCAT000000000000016","图像与多媒体",{"categoryNo":84,"categoryName":85,"taxonomyType":81,"taxonomyTypeDesc":82},[209,115,182,210,211],"resemble detect","https","identity",{"skillNo":213,"name":214,"nameZh":6,"description":215,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":199,"licenseFileUrl":216,"sourceHtmlUrl":217,"categories":218,"keywords":223,"publishedAt":42},"01M6AE30BSK2HCY4TQQ22Z8JRD","quality-playbook","Run a complete quality engineering audit on any codebase. Derives behavioral requirements from the code, generates spec-traced functional tests, runs a three-pass code review with regression tests, executes a multi-model spec audit (Council of Three), and produces a consolidated bug report with TDD-verified patches. Finds the 35% of real defects that structural code review alone cannot catch. Works with any language. Trigger on 'quality playbook', 'spec audit', 'Council of Three', 'fitness-to-purpose', or 'coverage theater'.","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fquality-playbook\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fgithub\u002Fawesome-copilot\u002Fblob\u002Fmain\u002Fskills\u002Fquality-playbook\u002FSKILL.md",[219,222],{"categoryNo":220,"categoryName":221,"taxonomyType":81,"taxonomyTypeDesc":82},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":84,"categoryName":85,"taxonomyType":81,"taxonomyTypeDesc":82},[224,225,226,227,146,92,9,114,228,229,210,230,231,232,233,234,235,236,237,238,239,240,99,241,21,242,243],"playbook","quality playbook","quality","code review","tdd","run","report","spec","bug report","council","engineering","functional","spec audit","bug","verified","multi-model","three","coverage","regression","structural"]