You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 

624 lines
21 KiB

{
"cells": [
{
"cell_type": "markdown",
"id": "06cf3063-9f3e-4551-a0d5-f08d9cabb927",
"metadata": {},
"source": [
"# Welcome to Week 2!\n",
"\n",
"## Frontier Model APIs\n",
"\n",
"In Week 1, we used multiple Frontier LLMs through their Chat UI, and we connected with the OpenAI's API.\n",
"\n",
"Today we'll connect with the APIs for Anthropic and Google, as well as OpenAI."
]
},
{
"cell_type": "markdown",
"id": "2b268b6e-0ba4-461e-af86-74a41f4d681f",
"metadata": {},
"source": [
"<table style=\"margin: 0; text-align: left;\">\n",
" <tr>\n",
" <td style=\"width: 150px; height: 150px; vertical-align: middle;\">\n",
" <img src=\"../important.jpg\" width=\"150\" height=\"150\" style=\"display: block;\" />\n",
" </td>\n",
" <td>\n",
" <h2 style=\"color:#900;\">Important Note - Please read me</h2>\n",
" <span style=\"color:#900;\">I'm continually improving these labs, adding more examples and exercises.\n",
" At the start of each week, it's worth checking you have the latest code.<br/>\n",
" First do a <a href=\"https://chatgpt.com/share/6734e705-3270-8012-a074-421661af6ba9\">git pull and merge your changes as needed</a>. Any problems? Try asking ChatGPT to clarify how to merge - or contact me!<br/><br/>\n",
" After you've pulled the code, from the llm_engineering directory, in an Anaconda prompt (PC) or Terminal (Mac), run:<br/>\n",
" <code>conda env update --f environment.yml --prune</code><br/>\n",
" Or if you used virtualenv rather than Anaconda, then run this from your activated environment in a Powershell (PC) or Terminal (Mac):<br/>\n",
" <code>pip install -r requirements.txt</code>\n",
" <br/>Then restart the kernel (Kernel menu >> Restart Kernel and Clear Outputs Of All Cells) to pick up the changes.\n",
" </span>\n",
" </td>\n",
" </tr>\n",
"</table>\n",
"<table style=\"margin: 0; text-align: left;\">\n",
" <tr>\n",
" <td style=\"width: 150px; height: 150px; vertical-align: middle;\">\n",
" <img src=\"../resources.jpg\" width=\"150\" height=\"150\" style=\"display: block;\" />\n",
" </td>\n",
" <td>\n",
" <h2 style=\"color:#f71;\">Reminder about the resources page</h2>\n",
" <span style=\"color:#f71;\">Here's a link to resources for the course. This includes links to all the slides.<br/>\n",
" <a href=\"https://edwarddonner.com/2024/11/13/llm-engineering-resources/\">https://edwarddonner.com/2024/11/13/llm-engineering-resources/</a><br/>\n",
" Please keep this bookmarked, and I'll continue to add more useful links there over time.\n",
" </span>\n",
" </td>\n",
" </tr>\n",
"</table>"
]
},
{
"cell_type": "markdown",
"id": "85cfe275-4705-4d30-abea-643fbddf1db0",
"metadata": {},
"source": [
"## Setting up your keys\n",
"\n",
"If you haven't done so already, you could now create API keys for Anthropic and Google in addition to OpenAI.\n",
"\n",
"**Please note:** if you'd prefer to avoid extra API costs, feel free to skip setting up Anthopic and Google! You can see me do it, and focus on OpenAI for the course. You could also substitute Anthropic and/or Google for Ollama, using the exercise you did in week 1.\n",
"\n",
"For OpenAI, visit https://openai.com/api/ \n",
"For Anthropic, visit https://console.anthropic.com/ \n",
"For Google, visit https://ai.google.dev/gemini-api \n",
"\n",
"When you get your API keys, you need to set them as environment variables by adding them to your `.env` file.\n",
"\n",
"```\n",
"OPENAI_API_KEY=xxxx\n",
"ANTHROPIC_API_KEY=xxxx\n",
"GOOGLE_API_KEY=xxxx\n",
"```\n",
"\n",
"Afterwards, you may need to restart the Jupyter Lab Kernel (the Python process that sits behind this notebook) via the Kernel menu, and then rerun the cells from the top."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "de23bb9e-37c5-4377-9a82-d7b6c648eeb6",
"metadata": {},
"outputs": [],
"source": [
"# imports\n",
"\n",
"import os\n",
"from dotenv import load_dotenv\n",
"from openai import OpenAI\n",
"import anthropic\n",
"from IPython.display import Markdown, display, update_display"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "f0a8ab2b-6134-4104-a1bc-c3cd7ea4cd36",
"metadata": {},
"outputs": [],
"source": [
"# import for google\n",
"# in rare cases, this seems to give an error on some systems, or even crashes the kernel\n",
"# If this happens to you, simply ignore this cell - I give an alternative approach for using Gemini later\n",
"\n",
"import google.generativeai"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "1179b4c5-cd1f-4131-a876-4c9f3f38d2ba",
"metadata": {},
"outputs": [],
"source": [
"# Load environment variables in a file called .env\n",
"# Print the key prefixes to help with any debugging\n",
"\n",
"load_dotenv()\n",
"openai_api_key = os.getenv('OPENAI_API_KEY')\n",
"anthropic_api_key = os.getenv('ANTHROPIC_API_KEY')\n",
"google_api_key = os.getenv('GOOGLE_API_KEY')\n",
"\n",
"if openai_api_key:\n",
" print(f\"OpenAI API Key exists and begins {openai_api_key[:8]}\")\n",
"else:\n",
" print(\"OpenAI API Key not set\")\n",
" \n",
"if anthropic_api_key:\n",
" print(f\"Anthropic API Key exists and begins {anthropic_api_key[:7]}\")\n",
"else:\n",
" print(\"Anthropic API Key not set\")\n",
"\n",
"if google_api_key:\n",
" print(f\"Google API Key exists and begins {google_api_key[:8]}\")\n",
"else:\n",
" print(\"Google API Key not set\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "797fe7b0-ad43-42d2-acf0-e4f309b112f0",
"metadata": {},
"outputs": [],
"source": [
"# Connect to OpenAI, Anthropic\n",
"\n",
"openai = OpenAI()\n",
"\n",
"claude = anthropic.Anthropic()"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "425ed580-808d-429b-85b0-6cba50ca1d0c",
"metadata": {},
"outputs": [],
"source": [
"# This is the set up code for Gemini\n",
"# Having problems with Google Gemini setup? Then just ignore this cell; when we use Gemini, I'll give you an alternative that bypasses this library altogether\n",
"\n",
"google.generativeai.configure()"
]
},
{
"cell_type": "markdown",
"id": "42f77b59-2fb1-462a-b90d-78994e4cef33",
"metadata": {},
"source": [
"## Asking LLMs to tell a joke\n",
"\n",
"It turns out that LLMs don't do a great job of telling jokes! Let's compare a few models.\n",
"Later we will be putting LLMs to better use!\n",
"\n",
"### What information is included in the API\n",
"\n",
"Typically we'll pass to the API:\n",
"- The name of the model that should be used\n",
"- A system message that gives overall context for the role the LLM is playing\n",
"- A user message that provides the actual prompt\n",
"\n",
"There are other parameters that can be used, including **temperature** which is typically between 0 and 1; higher for more random output; lower for more focused and deterministic."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "378a0296-59a2-45c6-82eb-941344d3eeff",
"metadata": {},
"outputs": [],
"source": [
"system_message = \"You are an assistant that is great at telling jokes\"\n",
"user_prompt = \"Tell a light-hearted joke for an audience of Data Scientists\""
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "f4d56a0f-2a3d-484d-9344-0efa6862aff4",
"metadata": {},
"outputs": [],
"source": [
"prompts = [\n",
" {\"role\": \"system\", \"content\": system_message},\n",
" {\"role\": \"user\", \"content\": user_prompt}\n",
" ]"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "3b3879b6-9a55-4fed-a18c-1ea2edfaf397",
"metadata": {},
"outputs": [],
"source": [
"# GPT-3.5-Turbo\n",
"\n",
"completion = openai.chat.completions.create(model='gpt-3.5-turbo', messages=prompts)\n",
"print(completion.choices[0].message.content)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "3d2d6beb-1b81-466f-8ed1-40bf51e7adbf",
"metadata": {},
"outputs": [],
"source": [
"# GPT-4o-mini\n",
"# Temperature setting controls creativity\n",
"\n",
"completion = openai.chat.completions.create(\n",
" model='gpt-4o-mini',\n",
" messages=prompts,\n",
" temperature=0.7\n",
")\n",
"print(completion.choices[0].message.content)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "f1f54beb-823f-4301-98cb-8b9a49f4ce26",
"metadata": {},
"outputs": [],
"source": [
"# GPT-4o\n",
"\n",
"completion = openai.chat.completions.create(\n",
" model='gpt-4o',\n",
" messages=prompts,\n",
" temperature=0.4\n",
")\n",
"print(completion.choices[0].message.content)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "1ecdb506-9f7c-4539-abae-0e78d7f31b76",
"metadata": {},
"outputs": [],
"source": [
"# Claude 3.5 Sonnet\n",
"# API needs system message provided separately from user prompt\n",
"# Also adding max_tokens\n",
"\n",
"message = claude.messages.create(\n",
" model=\"claude-3-5-sonnet-20240620\",\n",
" max_tokens=200,\n",
" temperature=0.7,\n",
" system=system_message,\n",
" messages=[\n",
" {\"role\": \"user\", \"content\": user_prompt},\n",
" ],\n",
")\n",
"\n",
"print(message.content[0].text)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "769c4017-4b3b-4e64-8da7-ef4dcbe3fd9f",
"metadata": {},
"outputs": [],
"source": [
"# Claude 3.5 Sonnet again\n",
"# Now let's add in streaming back results\n",
"\n",
"result = claude.messages.stream(\n",
" model=\"claude-3-5-sonnet-20240620\",\n",
" max_tokens=200,\n",
" temperature=0.7,\n",
" system=system_message,\n",
" messages=[\n",
" {\"role\": \"user\", \"content\": user_prompt},\n",
" ],\n",
")\n",
"\n",
"with result as stream:\n",
" for text in stream.text_stream:\n",
" print(text, end=\"\", flush=True)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "6df48ce5-70f8-4643-9a50-b0b5bfdb66ad",
"metadata": {},
"outputs": [],
"source": [
"# The API for Gemini has a slightly different structure.\n",
"# I've heard that on some PCs, this Gemini code causes the Kernel to crash.\n",
"# If that happens to you, please skip this cell and use the next cell instead - an alternative approach.\n",
"\n",
"gemini = google.generativeai.GenerativeModel(\n",
" model_name='gemini-1.5-flash',\n",
" system_instruction=system_message\n",
")\n",
"response = gemini.generate_content(user_prompt)\n",
"print(response.text)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "49009a30-037d-41c8-b874-127f61c4aa3a",
"metadata": {},
"outputs": [],
"source": [
"# As an alternative way to use Gemini that bypasses Google's python API library,\n",
"# Google has recently released new endpoints that means you can use Gemini via the client libraries for OpenAI!\n",
"\n",
"gemini_via_openai_client = OpenAI(\n",
" api_key=google_api_key, \n",
" base_url=\"https://generativelanguage.googleapis.com/v1beta/openai/\"\n",
")\n",
"\n",
"response = gemini_via_openai_client.chat.completions.create(\n",
" model=\"gemini-1.5-flash\",\n",
" messages=prompts\n",
")\n",
"print(response.choices[0].message.content)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "83ddb483-4f57-4668-aeea-2aade3a9e573",
"metadata": {},
"outputs": [],
"source": [
"# To be serious! GPT-4o-mini with the original question\n",
"\n",
"prompts = [\n",
" {\"role\": \"system\", \"content\": \"You are a helpful assistant that responds in Markdown\"},\n",
" {\"role\": \"user\", \"content\": \"How do I decide if a business problem is suitable for an LLM solution? Please respond in Markdown.\"}\n",
" ]"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "749f50ab-8ccd-4502-a521-895c3f0808a2",
"metadata": {},
"outputs": [],
"source": [
"# Have it stream back results in markdown\n",
"\n",
"stream = openai.chat.completions.create(\n",
" model='gpt-4o',\n",
" messages=prompts,\n",
" temperature=0.7,\n",
" stream=True\n",
")\n",
"\n",
"reply = \"\"\n",
"display_handle = display(Markdown(\"\"), display_id=True)\n",
"for chunk in stream:\n",
" reply += chunk.choices[0].delta.content or ''\n",
" reply = reply.replace(\"```\",\"\").replace(\"markdown\",\"\")\n",
" update_display(Markdown(reply), display_id=display_handle.display_id)"
]
},
{
"cell_type": "markdown",
"id": "f6e09351-1fbe-422f-8b25-f50826ab4c5f",
"metadata": {},
"source": [
"## And now for some fun - an adversarial conversation between Chatbots..\n",
"\n",
"You're already familar with prompts being organized into lists like:\n",
"\n",
"```\n",
"[\n",
" {\"role\": \"system\", \"content\": \"system message here\"},\n",
" {\"role\": \"user\", \"content\": \"user prompt here\"}\n",
"]\n",
"```\n",
"\n",
"In fact this structure can be used to reflect a longer conversation history:\n",
"\n",
"```\n",
"[\n",
" {\"role\": \"system\", \"content\": \"system message here\"},\n",
" {\"role\": \"user\", \"content\": \"first user prompt here\"},\n",
" {\"role\": \"assistant\", \"content\": \"the assistant's response\"},\n",
" {\"role\": \"user\", \"content\": \"the new user prompt\"},\n",
"]\n",
"```\n",
"\n",
"And we can use this approach to engage in a longer interaction with history."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "bcb54183-45d3-4d08-b5b6-55e380dfdf1b",
"metadata": {},
"outputs": [],
"source": [
"# Let's make a conversation between GPT-4o-mini and Claude-3-haiku\n",
"# We're using cheap versions of models so the costs will be minimal\n",
"\n",
"gpt_model = \"gpt-4o-mini\"\n",
"claude_model = \"claude-3-haiku-20240307\"\n",
"\n",
"gpt_system = \"You are a chatbot who is very argumentative; \\\n",
"you disagree with anything in the conversation and you challenge everything, in a snarky way.\"\n",
"\n",
"claude_system = \"You are a very polite, courteous chatbot. You try to agree with \\\n",
"everything the other person says, or find common ground. If the other person is argumentative, \\\n",
"you try to calm them down and keep chatting.\"\n",
"\n",
"gpt_messages = [\"Hi there\"]\n",
"claude_messages = [\"Hi\"]"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "1df47dc7-b445-4852-b21b-59f0e6c2030f",
"metadata": {},
"outputs": [],
"source": [
"def call_gpt():\n",
" messages = [{\"role\": \"system\", \"content\": gpt_system}]\n",
" for gpt, claude in zip(gpt_messages, claude_messages):\n",
" messages.append({\"role\": \"assistant\", \"content\": gpt})\n",
" messages.append({\"role\": \"user\", \"content\": claude})\n",
" completion = openai.chat.completions.create(\n",
" model=gpt_model,\n",
" messages=messages\n",
" )\n",
" return completion.choices[0].message.content"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "9dc6e913-02be-4eb6-9581-ad4b2cffa606",
"metadata": {},
"outputs": [],
"source": [
"call_gpt()"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "7d2ed227-48c9-4cad-b146-2c4ecbac9690",
"metadata": {},
"outputs": [],
"source": [
"def call_claude():\n",
" messages = []\n",
" for gpt, claude_message in zip(gpt_messages, claude_messages):\n",
" messages.append({\"role\": \"user\", \"content\": gpt})\n",
" messages.append({\"role\": \"assistant\", \"content\": claude_message})\n",
" messages.append({\"role\": \"user\", \"content\": gpt_messages[-1]})\n",
" message = claude.messages.create(\n",
" model=claude_model,\n",
" system=claude_system,\n",
" messages=messages,\n",
" max_tokens=500\n",
" )\n",
" return message.content[0].text"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "01395200-8ae9-41f8-9a04-701624d3fd26",
"metadata": {},
"outputs": [],
"source": [
"call_claude()"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "08c2279e-62b0-4671-9590-c82eb8d1e1ae",
"metadata": {},
"outputs": [],
"source": [
"call_gpt()"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "0275b97f-7f90-4696-bbf5-b6642bd53cbd",
"metadata": {},
"outputs": [],
"source": [
"gpt_messages = [\"Hi there\"]\n",
"claude_messages = [\"Hi\"]\n",
"\n",
"print(f\"GPT:\\n{gpt_messages[0]}\\n\")\n",
"print(f\"Claude:\\n{claude_messages[0]}\\n\")\n",
"\n",
"for i in range(5):\n",
" gpt_next = call_gpt()\n",
" print(f\"GPT:\\n{gpt_next}\\n\")\n",
" gpt_messages.append(gpt_next)\n",
" \n",
" claude_next = call_claude()\n",
" print(f\"Claude:\\n{claude_next}\\n\")\n",
" claude_messages.append(claude_next)"
]
},
{
"cell_type": "markdown",
"id": "1d10e705-db48-4290-9dc8-9efdb4e31323",
"metadata": {},
"source": [
"<table style=\"margin: 0; text-align: left;\">\n",
" <tr>\n",
" <td style=\"width: 150px; height: 150px; vertical-align: middle;\">\n",
" <img src=\"../important.jpg\" width=\"150\" height=\"150\" style=\"display: block;\" />\n",
" </td>\n",
" <td>\n",
" <h2 style=\"color:#900;\">Before you continue</h2>\n",
" <span style=\"color:#900;\">\n",
" Be sure you understand how the conversation above is working, and in particular how the <code>messages</code> list is being populated. Add print statements as needed. Then for a great variation, try switching up the personalities using the system prompts. Perhaps one can be pessimistic, and one optimistic?<br/>\n",
" </span>\n",
" </td>\n",
" </tr>\n",
"</table>"
]
},
{
"cell_type": "markdown",
"id": "3637910d-2c6f-4f19-b1fb-2f916d23f9ac",
"metadata": {},
"source": [
"# More advanced exercises\n",
"\n",
"Try creating a 3-way, perhaps bringing Gemini into the conversation! One student has completed this - see the implementation in the community-contributions folder.\n",
"\n",
"Try doing this yourself before you look at the solutions. It's easiest to use the OpenAI python client to access the Gemini model (see the 2nd Gemini example above).\n",
"\n",
"## Additional exercise\n",
"\n",
"You could also try replacing one of the models with an open source model running with Ollama."
]
},
{
"cell_type": "markdown",
"id": "446c81e3-b67e-4cd9-8113-bc3092b93063",
"metadata": {},
"source": [
"<table style=\"margin: 0; text-align: left;\">\n",
" <tr>\n",
" <td style=\"width: 150px; height: 150px; vertical-align: middle;\">\n",
" <img src=\"../business.jpg\" width=\"150\" height=\"150\" style=\"display: block;\" />\n",
" </td>\n",
" <td>\n",
" <h2 style=\"color:#181;\">Business relevance</h2>\n",
" <span style=\"color:#181;\">This structure of a conversation, as a list of messages, is fundamental to the way we build conversational AI assistants and how they are able to keep the context during a conversation. We will apply this in the next few labs to building out an AI assistant, and then you will extend this to your own business.</span>\n",
" </td>\n",
" </tr>\n",
"</table>"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "c23224f6-7008-44ed-a57f-718975f4e291",
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.11.11"
}
},
"nbformat": 4,
"nbformat_minor": 5
}