ramsey 0.0.1

A library for probabilistic models using Haiku and JAX

Ramsey

Probabilistic modelling using Haiku and JAX

About

Ramsey is a library for probabilistic models using Haiku and JAX. It builds upon the same module system that Haiku is using and is hence fully compatible with Haiku's and NumPyro's API. Ramsey implements (or rather intends to implement) neural and Gaussian process models and normalizing flows.

Installation

To install the latest GitHub release, just call the following on the command line:

pip install git+https://github.com/dirmeier/ramsey@v0.0.1

See also the installation instructions for Haiku and JAX.

Example usage

Ramsey uses to Haiku's module system to construct probabilistic models and define parameters. For instance, a simple neural process can be constructed like this:

import haiku as hk
import jax.random as random

from ramsey.data import sample_from_sinus_function
from ramsey.models import NP

def neural_process(**kwargs):
    dim = 128
    np = NP(
        decoder=hk.nets.MLP([dim] * 3 + [2]),
        latent_encoder=(
            hk.nets.MLP([dim] * 3), hk.nets.MLP([dim, dim * 2])
        )
    )
    return np(**kwargs)

key = random.PRNGKey(23)
(x, y), _ = sample_from_sinus_function(key)

neural_process = hk.transform(neural_process)
params = neural_process.init(key, x_context=x, y_context=y, x_target=x)

Why Ramsey

Just as other probabilistic languages are named after researchers in the field (e.g., Stan, Edward, Turing), Ramsey takes its name from one of my favourite philosophers/mathematicians, Frank Ramsey.

Author

Simon Dirmeier simon.dirmeier @ protonmail com