RAG example

Log in
RAG Indexing
RAG Query
LLM playground

Log in

This service is not intended to be shared by more than one user at a time! This restriction applies ONLY to the demo work of the service and does not affect the API and work in production. Therefore, all the /tick/ calls are not needed for production work.

Your name to avoid service sharing problems:

RAG indexing into a new database

Prepare data for indexing:

Data source files:

Data subfolder in "incoming" folder on server, a-zA-Z0-9-_, leave empty if not used:

Task ID, a-zA-Z0-9-_, leave empty for auto generate:

Indexing pipeline:

For QA pipelines file should be XLSX with columns: question answer. Data samples: Empty XLSX file sample, Real PSBank 534 QA RU, InsuranceQAv2 27988 QA EN,

Change existing database:

Add additional files to this existing index:

Data source files:

Remove files from index:

Enter original name of files to remove, one name in one line:

Remove task and ALL their files:

RAG query

You have to add documents to index first. Please use Indexing page to add documents to index.

Ask a question to RAG:

Question:

Optional params: use example of params for yagpt

RAG reply :

Here will be documents content

Here will be prompt

Ask many questions (slow):

File should be .XLSX with columns: question, filename, correct_answer:

File qoth questions:

LLM playground

Ask a question to LLM: Use sample for out-of-domain understanding, use more params

Choose model:

Service say 1:

User say 1:

Service say 2:

User say 2:

Service say 3:

User say 3:

Model reply: