Logging

Text models

You can start logging your OpenAI with very little change in your current code. We strongly suggest setting up logs to collect high quality data from your GPT4 requests.

Let us start with installing gigaml PyPi package. We currently have a python sdk, we are in the process of developing a typescript sdk.

python

    pip install gigaml

Here is how you can log the requests

python

  import openai, os
  from dotenv import load_dotenv
  # IMPORTING GIGA ML LIBRARY 
  from gigaml import OpenAI, GigaMlApi 

  load_dotenv('.env')

  client = OpenAI(
      api_key = os.environ.get('OPENAI_API_KEY'),

      #Initialise gigaml Client through API_KEY which can be generated via our console
      gigaml_client = GigaMlApi(token = os.environ.get("GIGAML_KEY")),
  )

  response = client.chat.completions.create(
      model = "gpt-3.5-turbo",
      messages = [{"role":"user", "content": "Who are you"}],
      tags = ["testing"],
      stream = True,
  )

  for chunk in response:
      print(chunk)

Vision Models

We allow for logging of requests and responses for vision models as well (ex. gpt-4-vision-preview). We accept both base_64 encoding and url in the ‘image_url’ field.

Logging vision models is the same as logging text models.

python

  import openai, os
  from dotenv import load_dotenv
  # IMPORTING GIGA ML LIBRARY 
  from gigaml import OpenAI, GigaMlApi 

  load_dotenv('.env')

  client = OpenAI(
      api_key = os.environ.get('OPENAI_API_KEY'),

      #Initialise gigaml Client through API_KEY which can be generated via our console
      gigaml_client = GigaMlApi(token = os.environ.get("GIGAML_KEY")),
  )

  response = client.chat.completions.create(
      model = "gpt-3.5-turbo",
      messages = [{"role":"user", "content": "Who are you"}],
      tags = ["testing"],
      stream = True,
  )

  for chunk in response:
      print(chunk)

When you log base_64 encoded images, we will not display the whole base_64 encoding in the dashboard. You will instead see something like this:

Download logs

You can easily download the logs from the dashboard. The logs downloaded will be compatible with OpenAI’s finetuning file format. You can select the logs you want to download and click on the download button on the top right corner.

Introduction

Datasets

Finetuning

Models

Pricing

Logging

Text models

Vision Models

Download logs

Introduction

Logging

Datasets

Finetuning

Models

Pricing

​Text models

​Vision Models

​Download logs

Text models

Vision Models

Download logs