Skip to content

Latest commit

 

History

History
22 lines (17 loc) · 1.38 KB

README.md

File metadata and controls

22 lines (17 loc) · 1.38 KB

NeuralLikelihoods

Source code for Neural Likelihoods via Cumulative Distribution Functions paper

Currently only the code for experiments/density/synthetic is working because the code is being rewritten from Tensorflow 1.x to Tensorflow 2.0.

Please create conda environment from provided environment.yml.

Training models in a notebook on simple generated data.

Open notebook and choose the model to train:

  1. experiments/density/synthetic/sin_normal/test.ipynb
  2. experiments/density/synthetic/sin_t/test.ipynb
  3. experiments/density/synthetic/inv_sin_normal/test.ipynb
  4. experiments/density/synthetic/inv_sin_t/test.ipynb
  5. experiments/density/synthetic/mv_nonlinear/test.ipynb

Running experiments:

  1. Experiment for each model/data set can be run by executing its script, for example experiments/density/synthetic/inv_sin_normal/monde_ar.py. User has to adjust configuration options that are set on the conf object in the experiment. These options depend on the hardware like available GPUs, available memory and number of CPUs.