AI Campus • Prayogika kalike
RAG mattu Vector Hudukaata Vyavasthegalu
Lexical retrieval mattu local vector search holisutta prati document na moolavannu pattehaccuva saulabhya ulisi.
Aarambhisuva modalu
- Python lists/dictionaries, document chunking basics mattu balakege anumati iruva chikka notes sangraha beku.
Ee oduva mattu planning abhyasagalige paid course athava API bekaagilla. Optional inference, training, IDE feature athava hosted sevege hardware, account athava paid access bekaagabahudu. Kharchu maaduva modalu local athava manual aayke balasi.
Official kalikeya maahitiyinda aarambhisi
Odalu free iruva moolagalannu 3 November 2026 randu parisheelisalagide. Documentation oduvudarinda provider athava subscription activate aaguvudilla.
Prayogika abhyasa maadi
1. Retrieval baseline kattiri
Sambandhita project guide nalliruva source-cited study-search starter anusarisiri. Aidu prashne baredu nireekshita source linegalannu dakhale maadi. Sangrahadalli uttara illada ondu prashne serisi.
2. Chunk mattu source ulisi
Chroma parichaya odi. Chunk ID, document ID, text mattu page/line metadata iruva recordgalannu roopisi. Embedding model athava chunking vidhaana badalaadaaga enannu marukattabeku endu plan maadi.
3. Top-k retrieval holisi
Optional aagi hondikeya local embedding model balasi local Chroma collection rachisi. Modala mooru resultgalannu baseline mattu kaiyaare gurutisida support passagegala jothe holisi. Generated answergalannu pratyeka extension aagi maatra evaluate maadi.
Kalitaddannu parikshe maadi
- Maralida chunk ID ninda ade moola text sigabeku.
- Aadhara illada prashnegalu test set nalli irabeku.
- Search result annu verified generated answer endu gurutisabaaradu.
Nimma karya, maadida parikshe mattu ulida mitigalannu sanna README nalli bareyiri. Kevala design maadida abhyasavannu execute maadida haage helabedi.