Hugging Face Transformers

Overview

This guide shows you how to integrate Latitude Telemetry into an existing application that uses Hugging Face Transformers for local model inference.

After completing these steps:

Every Transformers pipeline call can be captured as a log in Latitude.

Logs are grouped under a prompt, identified by a path, inside a Latitude project.

You can inspect inputs/outputs, measure latency, and debug Transformers-powered features from the Latitude dashboard.

You’ll keep calling Transformers exactly as you do today — Telemetry simply observes and enriches those calls.

Steps

Install requirements

Add the Latitude Telemetry package to your project:

pip install latitude-telemetry

uv add latitude-telemetry

poetry add latitude-telemetry

Wrap your Transformers-powered feature

Initialize Latitude Telemetry and wrap the code that calls Transformers using telemetry.capture.You can use the capture method as a decorator (recommended) or as a context manager:

Using decorator (recommended)

import os
from transformers import pipeline
from latitude_telemetry import Telemetry, Instrumentors, TelemetryOptions

telemetry = Telemetry(
    os.environ["LATITUDE_API_KEY"],
    TelemetryOptions(instrumentors=[Instrumentors.Transformers]),
)

@telemetry.capture(
    project_id=123,  # The ID of your project in Latitude
    path="generate-support-reply",  # Add a path to identify this prompt in Latitude
)
def generate_support_reply(input: str) -> str:
    generator = pipeline("text-generation", model="gpt2")
    result = generator(input, max_length=100)
    return result[0]["generated_text"]

Using context manager

import os
from transformers import pipeline
from latitude_telemetry import Telemetry, Instrumentors, TelemetryOptions

telemetry = Telemetry(
    os.environ["LATITUDE_API_KEY"],
    TelemetryOptions(instrumentors=[Instrumentors.Transformers]),
)

def generate_support_reply(input: str) -> str:
    with telemetry.capture(
        project_id=123,  # The ID of your project in Latitude
        path="generate-support-reply",  # Add a path to identify this prompt in Latitude
    ):
        generator = pipeline("text-generation", model="gpt2")
        result = generator(input, max_length=100)
        return result[0]["generated_text"]

The path:

Identifies the prompt in Latitude
Can be new or existing
Should not contain spaces or special characters (use letters, numbers, - _ / .)

Seeing your logs in Latitude

Once your feature is wrapped, logs will appear automatically.

Open the prompt in your Latitude dashboard (identified by path)

Go to the Traces section

Each execution will show:

Input and output messages
Model and token usage
Latency and errors
One trace per feature invocation

Each Transformers call appears as a child span under the captured prompt execution, giving you a full, end-to-end view of what happened.

Overview

Providers

Frameworks

Hugging Face Transformers

Overview

Requirements

Steps

Seeing your logs in Latitude

That’s it

​Overview

​Requirements

​Steps

​Seeing your logs in Latitude

​That’s it

Overview

Requirements

Steps

Seeing your logs in Latitude

That’s it