mxbai-embed-large-v1
Free.ai (self-hosted)
·
embeddings
·
~100 _Gün call
mxbai-embed-large-v1 an embedding model mixedbread.ai tarapyndan bina edildi. Semantic search, clustering, similarity.-de iň güýçli Free.ai GPUs üstünde öz-özüne-hosting — siziň gündelik token pool (100 tokens her jaň üçin) garşy azat işleýär. Apache 2.0 astynda çykaryldy - Free.ai astynda söwda üçin peýdalanyp bilner.
Köp soralýan soraglar
mxbai-embed-large-v1 metinni aňsat wektora (a float list) öwürýär. Onu semantik aramak, klasterleme, maslahat, aňsat aramak-artan emeli (RAG), we "bu metin şu metine meňzeşmi" meselesine degişli her bir iş üçin ullan.
Typical dimensions are 384, 768, 1024, or 1536 depending on the model. BGE-M3 emits 1024-dim; OpenAI Ada emits 1536. The API response includes the dimension so your vector DB picks the right index.
Modern embedding models (including most options on Free.ai) are trained on 100+ languages. Cross-language retrieval works — search in English, match documents in Spanish.
512den 8,192e çenli tokeni modele bagly. Uzak girdejiler kesilýär - uzun senedleri äparetlere bölmek üçin.
mxbai-embed-large-v1 öz GPUs-da işleýär we iň gymmatly gurallardan biri - gündelik azat pool-dan her bir çagyryş üçin ~100 tokeni çekýär. $5 = 200K tokeni.
Eý - POST /v1/embeddings/'e bir katyrlar sanawy we mxbai-embed-large-v1 bir wektorlar sanawy aňsatlyk bilen gaýtarýar. Bir soraga 2,048'e çenli baç ululyk.
L2-normalized by default — cosine similarity = dot product. Pass `normalize=false` if you want raw vectors for a different distance metric.
Her haýsy - Pinecone, Weaviate, Qdrant, Chroma, pgvector, FAISS, LanceDB. mxbai-embed-large-v1 gaýdyp getirýär aňsat JSON öçürmeler; DB hiç wagt model görmeýär.
Eý - POST to /v1/embeddings/ with model="mxbai-embed-large-v1". OpenAI-compatible response shape, so existing client libraries work unchanged. /api/ has the full reference.
Öz-özüne hosturlanan modeller metini GPU-da saklaýar we çagyryşdan soň ony boşadýar. Premium DPA bilen geçýär. Biz seniň girdişiň üstünde öwretmänis.
Öz-hosting-da gysga metin üçin sub-100ms, premium-da 100-500ms. Batch calls scale approximately linearly - 1,000 chunks complete in 2-10 seconds.
Eý - Free.ai embeddings-yň kommersiýa ulanmak hakyny berýär. Proýekt gözlegi, RAG borular, wektor-a-royalty-siz maslahat sistemalary bina et.