Browse Source

week 1

pull/37/head
mikeulator 5 months ago
parent
commit
0ca5d2181e
  1. 2
      week1/Guide to Jupyter.ipynb
  2. 4
      week1/day1.ipynb
  3. 2
      week1/day2 EXERCISE.ipynb
  4. 63
      week1/day2challenge.py
  5. 42
      week1/day5-llama3.2-result.md
  6. 32
      week1/day5-llama3.3-result.md
  7. 2
      week1/day5.ipynb
  8. 134
      week1/day5.py

2
week1/Guide to Jupyter.ipynb

@ -372,7 +372,7 @@
"name": "python", "name": "python",
"nbconvert_exporter": "python", "nbconvert_exporter": "python",
"pygments_lexer": "ipython3", "pygments_lexer": "ipython3",
"version": "3.11.11" "version": "3.11.9"
} }
}, },
"nbformat": 4, "nbformat": 4,

4
week1/day1.ipynb

@ -69,7 +69,7 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": null, "execution_count": 1,
"id": "4e2a9393-7767-488e-a8bf-27c12dca35bd", "id": "4e2a9393-7767-488e-a8bf-27c12dca35bd",
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
@ -559,7 +559,7 @@
"name": "python", "name": "python",
"nbconvert_exporter": "python", "nbconvert_exporter": "python",
"pygments_lexer": "ipython3", "pygments_lexer": "ipython3",
"version": "3.11.11" "version": "3.11.9"
} }
}, },
"nbformat": 4, "nbformat": 4,

2
week1/day2 EXERCISE.ipynb

@ -222,7 +222,7 @@
"name": "python", "name": "python",
"nbconvert_exporter": "python", "nbconvert_exporter": "python",
"pygments_lexer": "ipython3", "pygments_lexer": "ipython3",
"version": "3.11.11" "version": "3.11.9"
} }
}, },
"nbformat": 4, "nbformat": 4,

63
week1/day2challenge.py

@ -0,0 +1,63 @@
import ollama
import os
import requests
from bs4 import BeautifulSoup
from IPython.display import Markdown, display
MODEL = "llama3.2:3b-instruct-q8_0"
messages = [
{"role": "user", "content": "Describe some of the business applications of Generative AI"}
]
# response = ollama.chat(model=MODEL, messages=messages)
# print(response['message']['content'])
class Website:
def __init__(self, url):
"""
Create this Website object from the given url using the BeautifulSoup library
"""
self.url = url
response = requests.get(url)
soup = BeautifulSoup(response.content, 'html.parser')
self.title = soup.title.string if soup.title else "No title found"
try:
for irrelevant in soup.body(["script", "style", "img", "input"]):
irrelevant.decompose()
except:
pass
self.text = soup.body.get_text(separator="\n", strip=True)
system_prompt = "You are an assistant that analyzes the contents of a website \
and provides a short summary, ignoring text that might be navigation related. \
Respond in markdown."
def user_prompt_for(website):
user_prompt = f"You are looking at a website titled {website.title}"
user_prompt += "\nThe contents of this website is as follows; \
please provide a short summary of this website in markdown. \
If it includes news or announcements, then summarize these too.\n\n"
user_prompt += website.text
return user_prompt
def messages_for(website):
return [
{"role": "system", "content": system_prompt},
{"role": "user", "content": user_prompt_for(website)}
]
def summarize(url):
website = Website(url)
response = ollama.chat(
model = MODEL,
messages = messages_for(website)
)
return response['message']['content']
def display_summary(url):
summary = summarize(url)
print(summary)
display_summary("https://mike-tupper.com/")

42
week1/day5-llama3.2-result.md

@ -0,0 +1,42 @@
This appears to be a press release archive or news feed from Anthropic, a company that specializes in developing large language models and artificial intelligence. Here's a summary of the key points:
**Company Overview**
* Anthropic is a leader in developing AI models that can understand and generate human-like language.
* The company has developed several AI models, including Claude, which is a conversational AI model designed to assist with various tasks.
**Product Updates**
* Claude: A conversational AI model that can be used for a variety of tasks, including customer service, data enrichment, and more. New versions of Claude have been released regularly, including:
+ Claude 2 (available on Amazon Bedrock)
+ Claude 3 (features improved performance and new capabilities)
+ Claude 3 Haiku (a faster version of Claude 3)
+ Claude Pro (an enterprise-level version of Claude)
**Partnerships and Collaborations**
* Anthropic has partnered with several companies, including:
+ Amazon Web Services (AWS) to integrate Claude into AWS's services
+ Accenture to develop trusted solutions for enterprises
+ BCG (Boston Consulting Group) to expand access to safer AI
+ SKT (Samsung Electronics) on a partnership announcement
**Industry Recognition**
* Anthropic has received recognition from industry leaders and experts, including:
+ Dario Amodei's prepared remarks from the AI Safety Summit on Anthropic's Responsible Scaling Policy
+ Thoughts on the US Executive Order, G7 Code of Conduct, and Bletchley Park Summit by Anthropic
**Safety and Ethics**
* Anthropic has emphasized its commitment to safety and ethics in AI development, including:
+ Expanding access to safer AI with Amazon
+ Releasing a Responsible Scaling Policy
+ Aligning on child safety principles
**Other News**
* Anthropic has also made headlines for other reasons, such as:
+ Introducing Claude Android app
+ Fine-tuning Claude 3 Haiku in Amazon Bedrock
+ Evaluating prompts in the developer console

32
week1/day5-llama3.3-result.md

@ -0,0 +1,32 @@
Based on the provided webpage contents, here is a summary of the information:
**About Anthropic**
* Anthropic is a company that develops large-scale AI systems.
* Their research teams aim to create safer, steerable, and more reliable models.
**Careers**
* Anthropic has a careers page that lists open roles across various teams and offices.
* The company values direct evidence of ability, such as independent research, blog posts, or open-source software contributions.
* They do not require PhDs, degrees, or previous ML experience for technical staff positions.
* About half of the technical staff have a PhD, and about half had prior experience in ML.
* Anthropic sponsors visas and green cards for eligible candidates.
**Interview Process**
* Interviews are conducted over Google Meet, with a preference for PST office hours.
* Candidates can re-apply after 12 months if they are not successful initially.
* The company does not provide feedback on resumes or interviews.
**Remote Work**
* Anthropic staff typically come to the office regularly, but some may work remotely part-time or full-time.
* The company understands that moving can take time and offers a transitional phase for remote workers.
**Research**
* Anthropic's research teams focus on developing safer, steerable, and more reliable large-scale AI systems.
* The company is working at the frontier of AI research and development.
Overall, Anthropic appears to be a company that values innovation, expertise, and diversity in its workforce. They prioritize creating safe and reliable AI systems and offer opportunities for career growth and development.

2
week1/day5.ipynb

@ -475,7 +475,7 @@
"name": "python", "name": "python",
"nbconvert_exporter": "python", "nbconvert_exporter": "python",
"pygments_lexer": "ipython3", "pygments_lexer": "ipython3",
"version": "3.11.11" "version": "3.11.9"
} }
}, },
"nbformat": 4, "nbformat": 4,

134
week1/day5.py

@ -0,0 +1,134 @@
import ollama
import os
import requests
import json
from bs4 import BeautifulSoup
from IPython.display import Markdown, display
"""
Available Models:
llama3.3:latest a6eb4748fd29 42 GB 24 hours ago
granite3-moe:3b 157f538ae66e 2.1 GB 2 weeks ago
granite3-dense:8b 199456d876ee 4.9 GB 2 weeks ago
nemotron:70b-instruct-q5_K_M def2cefbe818 49 GB 6 weeks ago
llama3.2:3b-instruct-q8_0 e410b836fe61 3.4 GB 7 weeks ago
llama3.2:latest a80c4f17acd5 2.0 GB 2 months ago
reflection:latest 5084e77c1e10 39 GB 3 months ago
HammerAI/llama-3.1-storm:latest 876631929cf6 8.5 GB 3 months ago
granite-code:34b 4ce00960ca84 19 GB 3 months ago
llama3.1:8b 91ab477bec9d 4.7 GB 3 months ago
llama3.1-Q8-8b:latest 3d41179680d6 8.5 GB 3 months ago
nomic-embed-text:latest 0a109f422b47 274 MB 3 months ago
rjmalagon/gte-qwen2-7b-instruct-embed-f16:latest a94ce5b37c1c 15 GB 3 months ago
llama3:70b-instruct-q5_K_M 4e84a5514862 49 GB 3 months ago
llama3:8b 365c0bd3c000 4.7 GB 3 months ago
mistral-nemo:12b-instruct-2407-q8_0 b91eec34730f 13 GB 3 months ago
"""
MODEL = "llama3.3"
messages = [
{"role": "user", "content": "Describe some of the business applications of Generative AI"}
]
# response = ollama.chat(model=MODEL, messages=messages)
# print(response['message']['content'])
class Website:
"""
A utility class to represent a website that we have scraped, now with links
"""
url: str
title: str
body: str
links: list[str]
text: str
def __init__(self, url):
self.url = url
response = requests.get(url)
self.body = response.content
soup = BeautifulSoup(self.body, 'html.parser')
self.title = soup.title.string if soup.title else "No title found"
if soup.body:
try:
for irrelevant in soup.body(["script", "style", "img", "input"]):
irrelevant.decompose()
self.text = soup.body.get_text(separator="\n", strip=True)
except:
pass
else:
self.text = ""
links = [link.get('href') for link in soup.find_all('a')]
self.links = [link for link in links if link]
def get_contents(self):
return f"Webpage Title:\n{self.title}\nWebpage Contents:\n{self.text}\n\n"
link_system_prompt = "You are provided with a list of links found on a webpage. \
You are able to decide which of the links would be most relevant to include in a brochure about the company, \
such as links to an About page, or a Company page, or Careers/Jobs pages.\n"
link_system_prompt += "You should respond in JSON as in this example:"
link_system_prompt += """
{
"links": [
{"type": "about page", "url": "https://full.url/goes/here/about"},
{"type": "careers page": "url": "https://another.full.url/careers"}
]
}
"""
def get_links_user_prompt(website):
user_prompt = f"Here is the list of links on the website of {website.url} - "
user_prompt += "please decide which of these are relevant web links for a brochure about the company, respond with the full https URL in JSON format. \
Do not include Terms of Service, Privacy, email links.\n"
user_prompt += "Links (some might be relative links):\n"
user_prompt += "\n".join(website.links)
return user_prompt
def get_links(url):
website = Website(url)
response = ollama.chat(
model=MODEL,
messages=[
{"role": "system", "content": link_system_prompt},
{"role": "user", "content": get_links_user_prompt(website)}
],
format="json"
)
result = response['message']['content']
return json.loads(result)
def get_all_details(url):
result = "Landing page:\n"
result += Website(url).get_contents()
links = get_links(url)
# print("Found links:", links)
for link in links["links"]:
result += f"\n\n{link['type']}\n"
result += Website(link["url"]).get_contents()
return result
system_prompt = "You are an assistant that analyzes the contents of several relevant pages from a company website \
and creates a short professional sales brochure about the company for prospective customers, investors and recruits. Respond \
in markdown. Include details of company culture, customers and careers/jobs if you have the information."
def get_brochure_user_prompt(company_name, url):
user_prompt = f"You are looking at a company called: {company_name}\n"
user_prompt += f"Here are the contents of its landing page and other relevant pages; use this information to build a short brochure of the company in markdown.\n"
user_prompt += get_all_details(url)
user_prompt = user_prompt[:20000] # Truncate if more than 5,000 characters
return user_prompt
def create_brochure(company_name, url):
response = ollama.chat(
model=MODEL,
messages=[
{"role": "system", "content": system_prompt},
{"role": "user", "content": get_brochure_user_prompt(company_name, url)}
],
)
result = response['message']['content']
print(result)
create_brochure("Anthropic", "https://anthropic.com")
Loading…
Cancel
Save