Hugging Face's logo

Spaces:
Russellml
/
Omani-Chatbot
Sleeping

App Files Community
Omani-Chatbot / documentation /latency.md
Russell Jeffrey
Files
9532ac2
preview code
|
raw
history blame contribute delete
1.24 kB

A newer version of the Streamlit SDK is available: 1.52.1

Upgrade

Latency and Tracing Report

This project uses LangSmith to monitor latency, performance, and tracing of the chatbot workflow.
All executions—including STT, RAG retrieval, and response generation—are logged for review.

Accessing Reports

You can access detailed reports via the LangSmith dashboard:

👉 LangSmith Dashboard Link

Purpose

  • Latency Monitoring: Identify bottlenecks in STT, RAG retrieval, or LLM response generation.
  • Traceability: Full workflow visualization for debugging and compliance.
  • Accountability: Ensures the system meets the response-time requirements for mental health interventions.

How to Use

  • LangSmith web UI: After clicking on the link, go to the Examples tab.
  • Dataset Access: Due to limitations of the LangSmith web UI, we’ve exported the full run details to a CSV file, available in this repository:
    • 📂 Download Full Run Details (CSV).
    • Path: /data/langsmith-elile-bot-runs-2025-09-27.csv (relative to this repo’s root).

📂 Download Full Run Details (CSV)

Path: /data/langsmith-elile-bot-runs-2025-09-27.csv (relative to this repo’s root).