prysel-labs/multi-modal-rag-pipeline

About this asset

End-to-end vector embedding & ingestion pipeline for unstructured docs. Apache Beam + Python pipeline for document chunking and vector indexing.

Published by verified creator @prysel-labs for the Prysel ecosystem. Designed for real-time model integration, pipeline automation, and Prysel Chat assistance.

Key Capabilities & Applications

• Real-time Inference: Native streaming responses through Prysel Chat and Python SDK APIs.
• Secure Sandboxed Execution: Verified against the automated Prysel Trust & Safety rules engine.
• Modular Automation: Compatible with Prysel CLI jobs, Python client pipelines, and .NET services.

Integrate Multi-Modal-RAG-Pipeline into your Python scripts or notebooks using the official Prysel Python client (sdk/python/prysel):

Python 3 · prysel.client Copy

from prysel import PryselClient

# 1. Initialize Prysel client pointing to local cluster or cloud endpoint
client = PryselClient(
    base_url="https://market.prysel.com",
    token="<YOUR_PRYSEL_API_TOKEN>"
)

# 2. Allocate and plan runtime execution for Multi-Modal-RAG-Pipeline
plan = client.plan(
    model_ref="prysel-labs/multi-modal-rag-pipeline",
    model_source="marketplace",
    runtime="pytorch"
)

print(f"Asset provisioned successfully: {plan}")

Installation via pip:

Use the official high-performance .NET SDK (Prysel.Sdk) to automate deployments and workflows:

C# · Prysel.Sdk.PryselClient Copy

using System;
using System.Net.Http;
using System.Threading.Tasks;
using Prysel.Sdk;

// 1. Configure PryselClient with HttpClient
using var httpClient = new HttpClient { BaseAddress = new Uri("https://market.prysel.com") };
var client = new PryselClient(httpClient);

// 2. Authenticate session
var token = await client.LoginAsync("prysel-labs", "<your-password>");

// 3. Plan runtime deployment for Multi-Modal-RAG-Pipeline
var response = await client.PlanRuntimeAsync(new
{
    modelRef = "prysel-labs/multi-modal-rag-pipeline",
    runtime = "pytorch",
    hours = 1
});

Console.WriteLine($"Prysel deployment response: {response.StatusCode}");

Add NuGet Package:

Pull, push, and execute jobs with this asset directly using the Prysel CLI tools (docs/cli.md):

# 1. Log in to Prysel CLI
$ prysel login --username prysel-labs --password <your-password>

# 2. Pull dataset or model checkpoint
$ prysel dataset pull prysel-labs/multi-modal-rag-pipeline v1.2.0 ./checkpoints/multi-modal-rag-pipeline

# 3. Push local changes or updated weights
$ prysel model push prysel-labs/multi-modal-rag-pipeline ./checkpoints/multi-modal-rag-pipeline --tag v1.2.0 --framework pytorch

# 4. Run automated cloud evaluation or inference job
$ prysel job run Evaluation --input-json '{"model": "prysel-labs/multi-modal-rag-pipeline", "version": "v1.2.0"}'