reproduced_miwae / README.md
clementchadebec's picture
Update README.md
938be9a
metadata
language: en
tags:
  - pythae
  - reproducibility
license: apache-2.0

This model was trained with pythae. It can be downloaded or reloaded using the method load_from_hf_hub

>>> from pythae.models import AutoModel
>>> model = AutoModel.load_from_hf_hub(hf_hub_path="clementchadebec/reproduced_miwae")

Reproducibility

This trained model reproduces the results of the official implementation of [1].

Model Dataset Metric Obtained value Reference value
MIWAE (M=8, K=8) Dyn. Binarized MNIST NLL (5000 IS) 85.09 (0.00) 84.97 (0.10)

[1] Rainforth, Tom, et al. "Tighter variational bounds are not necessarily better." International Conference on Machine Learning. PMLR, 2018.