neuralsampler 0.0.10

neural sampler

NeuralSampler - Pytorch

Implementation of Neural Sampler.

Install

$ pip install neuralsampler

or install the latest version by

pip install -U git+https://JiahaoYao:{password}@github.com/JiahaoYao/neuralsampler.git@main

Install the jax (follow the official instruction here)

pip install jax jaxlib==0.1.64+[YOUR_CUDA_VERSION] -f https://storage.googleapis.com/jax-releases/jax_releases.html

e.g.

pip install --upgrade jax jaxlib==0.1.64+cuda101 -f https://storage.googleapis.com/jax-releases/jax_releases.html

Usage

import torch
from neuralsampler import neuralsampler

Run the code

python main.py 

Run the scripts

python scripts/test_job.py 

Demonstrations and tutorials

Link	Description
	Load our pretrained checkpoints and play with sampling, likelihood computation, and controllable synthesis (JAX + FLAX)
	Load our pretrained checkpoints and play with sampling, likelihood computation, and controllable synthesis (PyTorch)
	Tutorial of score-based generative models in JAX + FLAX
	Tutorial of score-based generative models in PyTorch

Ignore me (random things)

this repo is to collect all the random results and reproduce the experiments here 

and then jax 

i will use jax and flax, like shown here: https://github.com/yang-song/score_sde

there are the templates of building the jax neural networks (quite interesting to try this functional programming )

Todo list

• this library is on the pytorch
• i am also going to prepare the colab notebook
• the dataset is through the gdown: you can download the dataset from the google drive
• plz check the abf-mmd can reproduce the results! I have checked the codes are the same? I guessed the only issue might be just run enough runs! (Sun 09/05/2021 21:06)
• at least lots of things are connected now!
• update the colab module, going to download the dataset from the cloud!