llama-llm 0.0.10

Build on large language models faster

Llama

Stop prompt tuning. Create your own Generative AI.

Installation

    pip install llama-llm

Setup your keys

Go to powerml.co. Log in to get you API key and purchase credits.

Create ~/.powerml/configure_llama.yaml and put a key in it.

production:
    key: "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"

Try an example

Google Colab Example

Define the LLM interface

Define the input and output types. Don't forget the context. They help the LLM understand the data.

class Tweet(Type):
    tweet: str = Context("a viral tweet")
    likes: int = Context("likes the tweet gets")
    retweets: int = Context("retweets the tweet gets")


class User(Type):
    username: str = Context("a user's handle on twitter")

Instantiate an LLM engine

from llama import LLM

llm = LLM(name="tweets")

Use the LLM to generate something new

example_tweet = generate_tweet(user=User(username="lawrencekingyo"))
print("my first llama tweet!", example_tweet)

Train the LLM on your data

llm.add_data(data=get_tweet_data())

example_tweet = generate_tweet(user=User(username="lawrencekingyo"))

print("Tweet after adding data", example_tweet)

def get_tweet_data():
    return [
        [
            User(username="syswarren"),
            Tweet(
                tweet="Tools aren't going to make you great designers. Your way of thinking, attention to detail, and ability to see the bigger picture will.",
                likes=1000,
                retweets=81,
            ),
        ],
        [   User(username="TheJackForge"),
            Tweet(
                tweet="I don't like telling people how to live their lives, but you should probably learn how to use Figma.",
                likes=341,
                retweets=28,
            ),
        ],
        [
            User(username="iamharaldur"),
            Tweet(
                tweet="Remember when we had the mental energy to hate a new logo?",
                likes=1000,
                retweets=59,
            ),
        ],
        [
            User(username="lexfridman"),
            Tweet(
                tweet="ChatGPT puts a mirror to humanity.",
                likes=11100,
                retweets=874,
            ),
        ],
        [
            User(username="iamaaronwill"),
            Tweet(
                tweet="I had to make you uncomfortable otherwise you would never have moved. - The Universe",
                likes=4000,
                retweets=1000,
            ),
        ],
    ]

Look at the results, and add your feedback to improve the LLM

llm.improve(on="tweet", to="have more likes")
llm.improve(on="tweet", to="have over 100 retweets")
llm.improve(
    on="tweet",
    to="make it shorter",
    good_examples=[
        Tweet(
            tweet="Solopreneurs, don't chase more clients - it's a beast that'll destroy you. Aim for BETTER clients. Just 2-3 great ones can set you for life. Want to know why and what to do? Let me tell you. #smallbusiness",
            likes=45,
            retweets=10,
        )
    ],
    bad_examples=[
        Tweet(
            tweet="They tell you to chase more clients. If you're a Solopreneur providing a professional service, you're feeding a beast that will destroy you. Your goal is not MORE clients. Your goal is BETTER clients. 2-3 great clients could set you for life. I'll tell you why and what to do.",
            likes=5,
            retweets=1,
        )
    ],
)
llm.improve(on="tweet", to="have no hashtags")

tweets = [generate_tweet(user=User(username="lawrencekingyo")) for i in range(5)]

print("tweets: ", [str(tweet) for tweet in tweets])

Build complex programs