Closed
Milestone
Version 2.1
Version 2.1 will contain a series of important added functionalities:
- Make a custom
Machine
fully implementable in pure Python; - Make a custom
Sampler
fully implementable in pure Python; - Add support for PyTorch-based models;
- Add density matrices;
- Implement all high-order applications (Supervised, Unsupervised, Vmc) in pure python
Unstarted Issues (open and unassigned)
0
Ongoing Issues (open and assigned)
0
Completed Issues (closed)
11
- 2.1: batch_size as machine option
- 2.1: Rename batch_size
- 2.1: Drop support for python 2.x
- 2.1: Remove c++ driver classes for high-order methods
- 2.1: float types for covariance_sv
- 2.1: Change options names in sr
- 2.1: Jax machines issues on Travis
- 2.1: Structure of generic MCMC samplers
- 2.1 : The current interface of pure-python machines is inconsistent
- Create a Python-based converter between PyTorch models and NetKet Machine
- Make Machine class (and maybe others) derivable in python
Loading
Loading
Loading