Skip to content

Commit

Permalink
website
Browse files Browse the repository at this point in the history
  • Loading branch information
@mlelarge
mlelarge committed Jan 20, 2025
1 parent 434941e commit 33b3c70
Showing 1 changed file with 255 additions and 0 deletions.
255 changes: 255 additions & 0 deletions llm/website_ollama.ipynb
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,255 @@
{
"cells": [
{
"cell_type": "markdown",
"id": "d465baef-080f-4cdb-b1c3-465e4e1ce305",
"metadata": {},
"source": [
"# Running a local LLM with ollama\n",
"\n",
"[Ollama](https://ollama.com/) is probably the easiest way to run a LLM on your local machine.\n",
"\n",
"To run the code of this notebook on your machine, you will need [ollama-python](https://github.com/ollama/ollama-python).\n",
"\n",
"If you want to run this notebook on Google Colab, you can follow the instructions below."
]
},
{
"cell_type": "markdown",
"id": "720847c4-1c9a-4ea9-8de5-2a0d198a4af8",
"metadata": {},
"source": [
"## Instructions to run Ollama on Colab:\n",
"- first install xterm\n",
"```\n",
"!pip install colab-xterm #https://pypi.org/project/colab-xterm/\n",
"%load_ext colabxterm\n",
"```\n",
"- start a xterm and inside the terminal start an ollama server:\n",
"```\n",
"%xterm\n",
"# curl https://ollama.ai/install.sh | sh\n",
" # ollama serve & ollama pull mistral\n",
"```\n",
"- in a cell, check that you have your model\n",
"```\n",
"!ollama list\n",
"```\n",
"- in a cell, install ollama-python\n",
"```\n",
"!pip install ollama\n",
"```"
]
},
{
"cell_type": "markdown",
"id": "13e9b490-6c0a-4563-bbf7-03c2bb268449",
"metadata": {},
"source": [
"## Check that everything is running fine"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "dd16d74c-ed5e-4826-952e-f1809e48fae6",
"metadata": {},
"outputs": [],
"source": [
"from ollama import chat\n",
"from ollama import ChatResponse\n",
"from IPython import display"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "1f804638-4101-46d4-86eb-c58e4c005394",
"metadata": {},
"outputs": [],
"source": [
"response: ChatResponse = chat(model='mistral', messages=[\n",
" {\n",
" 'role': 'user',\n",
" 'content': 'Why is the sky blue?',\n",
" },\n",
"])\n",
"\n",
"print(response.message.content)"
]
},
{
"cell_type": "markdown",
"id": "0c81abfc-3a45-4d7a-ae0c-ad2f73f63783",
"metadata": {},
"source": [
"## Choose your model (you will need to pull it before) and make a nice website!"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "fd80bb2c-d17f-4957-93ea-4abf89a1579f",
"metadata": {},
"outputs": [],
"source": [
"model = 'mistral'"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "99eb2596-ad7a-4e12-9dab-957c4d2e3819",
"metadata": {},
"outputs": [],
"source": [
"prompt_0 = \"Write a webpage for the course: Large Language Models : Introduction and Applications for Code given by Nathanaël FIJALKOW and Marc LELARGE in the Master MVA. I will pipe your output directly to a file, so just give me ```html\\n[page here]\\n```.\""
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "1e4c077a-fd5c-4f20-b1df-c33d506ec61c",
"metadata": {},
"outputs": [],
"source": [
"response_0: ChatResponse = chat(model=model, messages=[\n",
" {\n",
" 'role': 'user',\n",
" 'content': prompt_0,\n",
" },\n",
"])"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "ccef4c65-7df9-45dc-95eb-4fe001f46177",
"metadata": {},
"outputs": [],
"source": [
"print(response_0.message.content)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "43f79f75-f0c9-40c6-8686-7aedde090f12",
"metadata": {},
"outputs": [],
"source": [
"page = response_0.message.content.split(\"```\")[1].partition(\"\\n\")[2]"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "a729c73f-1b21-475f-b381-bd25fc2158c6",
"metadata": {},
"outputs": [],
"source": [
"display.HTML(page)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "6ad3fa84-d37b-45fc-95e2-afaf8618fc5f",
"metadata": {},
"outputs": [],
"source": [
"def parse_html(response):\n",
" return response.message.content.split(\"```\")[1].partition(\"\\n\")[2]\n",
"\n",
"def write_html_file(file_name, content):\n",
" try:\n",
" with open(file_name, 'w') as file:\n",
" file.write(content)\n",
" print(f\"HTML file '{file_name}' created successfully.\")\n",
" except Exception as e:\n",
" print(f\"An error occurred while writing the HTML file: {e}\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "5446b7bf-67c7-4de2-a06e-543bc93af478",
"metadata": {},
"outputs": [],
"source": [
"write_html_file('webpage_mistral_v0.html', page)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "60eddf60-b0ba-4f3c-a977-c916a9706e91",
"metadata": {},
"outputs": [],
"source": [
"prompt_1 = \"Add more detail and better html and css. I will pipe your output directly to a file, so just give me ```html\\n[page here]\\n``` so that it fits in one file.\""
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "b3e07184-84ae-48c9-ac04-4c912bbeb484",
"metadata": {},
"outputs": [],
"source": [
"response_1: ChatResponse = chat(model=model, messages=[\n",
" {\n",
" 'role': 'user',\n",
" 'content': prompt_0,\n",
" },\n",
" response_0.message,\n",
" {\n",
" 'role': 'user',\n",
" 'content': prompt_1\n",
" },\n",
"])"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "fe439346-e76e-4381-a7d5-df88472203c8",
"metadata": {},
"outputs": [],
"source": [
"new_page = parse_html(response_1)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "5360dbf9-7782-4b90-845e-ec759ca8a9ea",
"metadata": {},
"outputs": [],
"source": [
"write_html_file('webpage_mistral_v1.html', new_page)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "chess",
"language": "python",
"name": "chess"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.12.4"
}
},
"nbformat": 4,
"nbformat_minor": 5
}

0 comments on commit 33b3c70

Please sign in to comment.