outlines: A Python repository from rishiraj

Multiple choices

You can reduce the completion to a choice between multiple possibilities:

import outlines

model = outlines.models.transformers("mistralai/Mistral-7B-Instruct-v0.2")

prompt = """You are a sentiment-labelling assistant.
Is the following review positive or negative?

Review: This restaurant is just awesome!
"""

generator = outlines.generate.choice(model, ["Positive", "Negative"])
answer = generator(prompt)

Type constraint

You can instruct the model to only return integers or floats:

import outlines

model = outlines.models.transformers("WizardLM/WizardMath-7B-V1.1")

prompt = "<s>result of 9 + 9 = 18</s><s>result of 1 + 2 = "
answer = outlines.generate.format(model, int)(prompt)
print(answer)
# 3

prompt = "sqrt(2)="
generator = outlines.generate.format(model, float)
answer = generator(prompt, max_tokens=10)
print(answer)
# 1.41421356

Efficient regex-structured generation

Outlines also comes with fast regex-structured generation. In fact, the choice and format functions above all use regex-structured generation under the hood:

import outlines

model = outlines.models.transformers("mistralai/Mistral-7B-Instruct-v0.2")

prompt = "What is the IP address of the Google DNS servers? "

generator = outlines.generate.text(model)
unstructured = generator(prompt, max_tokens=30)

generator = outlines.generate.regex(
    model,
    r"((25[0-5]|2[0-4]\d|[01]?\d\d?)\.){3}(25[0-5]|2[0-4]\d|[01]?\d\d?)",
)
structured = generator(prompt, max_tokens=30)

print(unstructured)
# What is the IP address of the Google DNS servers?
#
# Passive DNS servers are at DNS servers that are private.
# In other words, both IP servers are private. The database
# does not contain Chelsea Manning

print(structured)
# What is the IP address of the Google DNS servers?
# 2.2.6.1

Unlike other libraries, regex-structured generation in Outlines is almost as fast as non-structured generation.

Efficient JSON generation following a Pydantic model

Outlines 〰 allows to guide the generation process so the output is guaranteed to follow a JSON schema or Pydantic model:

from enum import Enum
from pydantic import BaseModel, constr

import outlines
import torch


class Weapon(str, Enum):
    sword = "sword"
    axe = "axe"
    mace = "mace"
    spear = "spear"
    bow = "bow"
    crossbow = "crossbow"


class Armor(str, Enum):
    leather = "leather"
    chainmail = "chainmail"
    plate = "plate"


class Character(BaseModel):
    name: constr(max_length=10)
    age: int
    armor: Armor
    weapon: Weapon
    strength: int


model = outlines.models.transformers("mistralai/Mistral-7B-Instruct-v0.2")

# Construct structured sequence generator
generator = outlines.generate.json(model, Character)

# Draw a sample
rng = torch.Generator(device="cuda")
rng.manual_seed(789001)

character = generator("Give me a character description", rng=rng)

print(repr(character))
# Character(name='Anderson', age=28, armor=<Armor.chainmail: 'chainmail'>, weapon=<Weapon.sword: 'sword'>, strength=8)

character = generator("Give me an interesting character description", rng=rng)

print(repr(character))
# Character(name='Vivian Thr', age=44, armor=<Armor.plate: 'plate'>, weapon=<Weapon.crossbow: 'crossbow'>, strength=125)

The method works with union types, optional types, arrays, nested schemas, etc. Some field constraints are not supported yet, but everything else should work.

Efficient JSON generation following a JSON Schema

Sometimes you just want to be able to pass a JSON Schema instead of a Pydantic model. We've got you covered:

import outlines

schema = '''{
    "title": "Character",
    "type": "object",
    "properties": {
        "name": {
            "title": "Name",
            "maxLength": 10,
            "type": "string"
        },
        "age": {
            "title": "Age",
            "type": "integer"
        },
        "armor": {"$ref": "#/definitions/Armor"},
        "weapon": {"$ref": "#/definitions/Weapon"},
        "strength": {
            "title": "Strength",
            "type": "integer"
        }
    },
    "required": ["name", "age", "armor", "weapon", "strength"],
    "definitions": {
        "Armor": {
            "title": "Armor",
            "description": "An enumeration.",
            "enum": ["leather", "chainmail", "plate"],
            "type": "string"
        },
        "Weapon": {
            "title": "Weapon",
            "description": "An enumeration.",
            "enum": ["sword", "axe", "mace", "spear", "bow", "crossbow"],
            "type": "string"
        }
    }
}'''

model = outlines.models.transformers("mistralai/Mistral-7B-Instruct-v0.2")
generator = outlines.generate.json(model, schema)
character = generator("Give me a character description")

Using context-free grammars to guide generation

Formal grammars rule the world, and Outlines makes them rule LLMs too. You can pass any context-free grammar in the EBNF format and Outlines will generate an output that is valid to this grammar:

import outlines

arithmetic_grammar = """
    ?start: expression

    ?expression: term (("+" | "-") term)*

    ?term: factor (("*" | "/") factor)*

    ?factor: NUMBER
           | "-" factor
           | "(" expression ")"

    %import common.NUMBER
"""

model = outlines.models.transformers("WizardLM/WizardMath-7B-V1.1")
generator = outlines.generate.cfg(model, arithmetic_grammar)
sequence = generator("Alice had 4 apples and Bob ate 2. Write an expression for Alice's apples:")

print(sequence)
# (8-2)

This was a very simple grammar, and you can use outlines.generate.cfg to generate syntactically valid Python, SQL, and much more than this. Any kind of structured text, really. All you have to do is search for "X EBNF grammar" on the web, and take a look at the Outlines grammars module.

Open functions

Outlines can infer the structure of the output from the signature of a function. The result is a dictionary, and can be passed directly to the function using the usual dictionary expansion syntax **:

import outlines


def add(a: int, b: int):
    return a + b

model = outlines.models.transformers("WizardLM/WizardMath-7B-V1.1")
generator = outlines.generate.json(model, add)
result = generator("Return json with two integers named a and b respectively. a is odd and b even.")

print(add(**result))
# 3

A great advantage of passing functions directly to specify the structure is that the structure of the LLM will change with the function's definition. No need to change the code at several places!

Prompting

Building prompts can get messy. Outlines makes it easier to write and manage prompts by encapsulating templates inside "template functions".

These functions make it possible to neatly separate the prompt logic from the general program logic; they can be imported from other modules and libraries.

Template functions require no superfluous abstraction, they use the Jinja2 templating engine to help build complex prompts in a concise manner:

import outlines

examples = [
    ("The food was disgusting", "Negative"),
    ("We had a fantastic night", "Positive"),
    ("Recommended", "Positive"),
    ("The waiter was rude", "Negative")
]

@outlines.prompt
def labelling(to_label, examples):
    """You are a sentiment-labelling assistant.

    {% for example in examples %}
    {{ example[0] }} // {{ example[1] }}
    {% endfor %}
    {{ to_label }} //
    """

model = outlines.models.transformers("mistralai/Mistral-7B-Instruct-v0.2")
prompt = labelling("Just awesome", examples)
answer = outlines.generate.text(model)(prompt, max_tokens=100)

rishiraj/outlines