OpenAI API Mastery with Python: From Zero to Expert

Explore the ChatGPT interface, try built-in prompts, and craft your own prompts to generate ideas—from indoor dog activities to employee prompts—while learning how ChatGPT works.

Explore ChatGPT UI features, compare GPT 3.5 and GPT four, apply custom instructions, adjust settings and themes, manage data controls, shared chats, and export data.

Learn prompt design and engineering to guide ChatGPT and OpenAI models with zero-shot, one-shot, and few-shot learning for accurate, relevant outputs.

Compare good prompts with bad prompts to show how specificity and clear direction control GPT outputs, illustrated by a dog’s day scenario from wake to bed.

Learn how short learning via prompts adapts pre-trained OpenAI models to specific tasks, using zero-shot, one-shot, and few-shot examples to shape responses.

Explore zero shot, one shot, and few shot learning prompts with ChatGPT through concrete movie recommendation examples, and learn how prompts shape outputs.

Learn to build a tweet sentiment classifier with ChatGPT using zero-shot and few-shot learning, designing prompts and templates that produce labeled outputs like neutral, positive, or negative.

Learn to design prompts for a tweet sentiment classifier using few-shot learning, with examples and instructions. Adapt prompts to output only sentiments (positive, negative, neutral) and ensure consistent formatting.

Master prompt design and engineering with a defined prompt, applying five essential principles—clarity, conciseness, focus, and consistency—and letting ChatGPT act as someone proven to be powerful to elicit precise responses.

Explore essential large language model concepts, transformer architectures, tokenization, and the OpenAI Python library; learn to create, send, and interpret API responses while managing costs for GPT development.

Explore the foundations of large language models and natural language processing, including AI, ML, deep learning, and transformers, and examine tasks like text classification, translation, question answering, and text generation.

Explore OpenAI APIs and their models via API requests, including GPT-4, GPT-3.5, GPT-2, DALL‑E, text to speech, and Whisper, with pricing per thousand tokens.

Test OpenAI language models in the playground, logging in and selecting prompts. Understand pricing, free $5 credit, token-based costs, and rate limits for API usage.

Explore testing OpenAI models in the playground without coding, learning how completion works, and adjusting temperature and max tokens while exploring model options and the code view for Python API.

Explore the chat endpoint in OpenAI API mastery, learn to use system prompts to control model behavior, switch from completion to chat, adjust temperature, and observe usage metrics.

Explore building an OpenAI assistant in your app with the assistant API, defining instructions, selecting models, and using tools like code interpreter, retrieval, and function calling with PDFs and CSVs.

Learn to enhance an assistant using custom data with PDF and CSV uploads, leveraging retrieval and code interpreter tools to query, visualize, and analyze information.

Learn to analyze a csv file with the OpenAI code interpreter: upload the csv, inspect columns and records, compute statistics, and visualize distributions using Python and pandas.

Learn to create and securely save your OpenAI API key by phone verification and generating a secret key in the API key tab for safe Python use.

Set up Python for OpenAI by configuring API keys, creating a virtual environment, and installing the OpenAI library, then test access to OpenAI models hosted on OpenAI servers.

learn to store your api key in a toml config file instead of environment variables, load it with tomllib in python, and prepare it for OpenAI requests in web apps.

Learn to send your first OpenAI API post request in Python using the chat completion flow, including system and user messages, model selection, and API key configuration.

Define a chart completion with the OpenAI api, set the model to GPT 3.5 turbo, build system and user messages, and inspect the chat response.

Learn to send an OpenAI API request in Python using PyCharm, inspect chat responses, and navigate choices, messages, and usage tokens for the GPT 3.5 turbo 0125 model.

Learn to stream OpenAI's chat completion API in Python by enabling stream=true, processing responses in chunks, and extracting delta.content for progressive output.

Develop a ChatGPT app with OpenAI in Streamlit using chat input and chat message elements; learn to display user and assistant messages, install Streamlit, and run the app.

Build a Streamlit chat app that mirrors ChatGPT using the OpenAI chat completion API, including session state for message history and a two-panel layout with input and messages.

Initialize the OpenAI API key and config, build a chat completion API call, and display streaming responses in a Streamlit chat app with session history.

Explore the assistant api beyond the chat completion api, unlocking persistent threading along with knowledge retrieval, code interpreter, and function calling to build a robust ai assistant.

Learn how the assistant API powers back-end operations, enabling parallel tools, function calling, and access to GPT-3.5, GPT-4 (premium), whisper, DALL·E, and persistent threading.

Review the assistant api playground to understand backend operations, including creating a thread, sending messages, running tasks in python, and tracking the run status to retrieve responses.

Explore the run lifecycle of the assistant API, distinguishing terminal states (completed, failed, canceled, expired) from intermediate states (queued, in progress, required action), including external functions triggering actions.

Create an assistant in Python by wiring the assistant, thread, and run objects. Upload CSV knowledge to storage, attach file IDs, and enable code interpreter and knowledge retrieval tools.

Create and configure a conversation thread and its messages to store user and assistant exchanges using client.beta.threads.create, then link messages by thread id with content and role.

Create a run by providing thread id and assistant id, then monitor the run lifecycle from queued to completed, retrieve the assistant's message from the thread, and view the response.

Create and manage an OpenAI assistant workflow by building messages, threads, and runs, then loop to send prompts and fetch completed responses.

visualize data with a csv file using a data visualizer, plotting mpg vs hp, while handling text and image responses and debugging run statuses and threads.

Validate assistant responses by checking content types, handling text or image data, downloading image files via ids, displaying visuals with Pillow, and building a Streamlit chat app.

Create a data visualizer assistant app in Streamlit, using the OpenAI assistant API with threads and messages, and a run process to display chat responses in the app.

Initialize the streamlit session state with messages and a thread id, then create an empty thread with the client beta threads create, and display the assistant and thread ids.

Display chat history stored in session state by iterating messages and rendering each with st.chat_message. Render roles and content to distinguish user and assistant messages in a sample Streamlit app.

Send prompt to the assistant with assistant ID and thread ID, validate the response as text or image, and display text in chat or image in streamlit while updating messages.

Rerun the data visualizer app in Streamlit to explore an uploaded dataset, view 32 records and features, and generate interactive plots like scatter, bar, box, and violin plots.

Explore how fine tuning an OpenAI model compares to prompt engineering, and learn when to use fine tuning to improve performance on complex tasks and specific applications.

Explore how fine tuning a pre-trained GPT model with specialized linear algebra data shapes weights and biases, enhancing domain-specific answers while preserving general question answering.

Learn the four-step fine-tuning workflow for OpenAI models, including data preparation in a specific JSON format, model selection, training, evaluation, and deploying the fine-tuned model under tier requirements.

Explore how to view and manage API limits, upgrade to tier one and other tiers, and fine-tune models like Babbage and Davinci, with insights on rate limits and billing.

Prepare data for fine tuning an OpenAI model using the CQ science dataset, turning questions, correct answers, and supporting text into prompt and completion pairs for Babbage 002 and Davinci.

Prepare data for OpenAI fine tune by building a Python script that reads train and test csv files with pandas, drops empty records, and validates the data shapes.

Format and prepare fine-tuning data by pairing prompts with ideal completions, including answer and support, then build train and test jsonl files and save for OpenAI model training.

Train OpenAI models through fine tuning using prepared jsonl data. Upload train and test data, select Babbage 002 or GPT 3.5 turbo, and monitor training and validation loss.