Name: tessl/pypi-pymc3
Rating: 0.68 (1 reviews)
Author: tessl

Blog Docs Log in Get started

tessl/pypi-pymc3

Probabilistic Programming in Python: Bayesian Modeling and Probabilistic Machine Learning with Theano

0.94x

Overview

Eval results

Files

Evaluation results

45%

Marketing Causal Interventions

Causal graph interventions/observations and model graph exports

Criteria

Without context

With context

Model structure

66%

100%

Posterior fit

100%

Do intervention

20%

Observation update

25%

20%

Predictive draws

50%

40%

Graph exports

20%

Seed control

100%

80%

70%

Labeled Sales Forecasting

Labeled dims and data binding via coords/Data/Minibatch

Criteria

Without context

With context

Coords Model

100%

Shared Data

100%

Data Updates

100%

Minibatch Inputs

Loglik Scaling

Posterior Labels

100%

96%

-3%

Campaign Response Model

Model specification with distributions, deterministics, and potentials

Criteria

Without context

With context

Model setup

93%

86%

Priors & offsets

100%

Logistic rates

100%

90%

Deterministic summaries

100%

Observed likelihood

100%

Penalty potential

100%

15%

Log Probability Helpers

Log-probability and math helpers for PyTensor-backed graphs

Criteria

Without context

With context

Graph distribution

Log density via logp

Tail probability

Quantile via icdf

Tensor outputs

100%

90%

Prior Predictive and Posterior Forecasting for Daily Counts

Prior and posterior predictive simulation utilities

Criteria

Without context

With context

Model structure

100%

Prior predictive

100%

Posterior sampling

100%

Posterior predictive

60%

Reproducibility & summaries

80%

100%

5%

-49%

Blocked Sampler Configuration

Custom step-method blocking and proposal configuration

Criteria

Without context

With context

Model wiring

32%

Manual blocking

80%

Proposal selection

25%

Deterministic runs

100%

33%

Acceptance reporting

40%

91%

Variational Logistic Regression

Variational inference APIs (ADVI, SVGD, KLqp, custom approximations)

Criteria

Without context

With context

Modeling with PyMC

75%

80%

Mean-field VI

100%

Full-rank VI

100%

Seed control

80%

100%

Posterior predictive

72%

80%

70%

Logistic ODE Inference

ODE-integrated probabilistic models via DifferentialEquation

Criteria

Without context

With context

DifferentialEquation

100%

Latent priors

100%

Observable data

33%

46%

MCMC sampling

100%

Posterior predictive

Predictive outputs

53%

100%

HSGP 1D Regression

Gaussian process modeling classes and HSGP approximations

Criteria

Without context

With context

HSGP model

100%

Positive priors

100%

Latent prior

100%

Noise likelihood

100%

Posterior sampling

100%

Posterior predictive

100%

Prior predictive

100%

Automatic MCMC Coin Bias

MCMC sampling with auto step selection (NUTS, HMC, Metropolis, Slice, Gibbs)

Criteria

Without context

With context

Model setup

100%

Auto sampler

100%

Seed reuse

100%

Posterior array

100%

HDI via package

100%

Install with Tessl CLI

npx tessl i tessl/pypi-pymc3

Evaluated: about 2 months ago
Agent: Claude Code

Table of Contents

Marketing Causal Interventions Labeled Sales Forecasting Campaign Response Model Log Probability Helpers Prior Predictive and Posterior Forecasting for Daily Counts Blocked Sampler Configuration Variational Logistic Regression Logistic ODE Inference HSGP 1D Regression Automatic MCMC Coin Bias