nutorbit / SAC

A PyTorch implementation of the SAC paper

SAC

This is a PyTorch implementation of Soft Actor Critic, including Impala Architecture.

Running the code

training a model.

$ python train.py

testing a model.

$ python test.py PATH

Where PATH is the path to your model.

show the logs.

$ tensorboard --logdir LOG --port 6006

Where LOG is the path of your logs, And display on localhost:6006/

To-do list

Here's my checklist:

• Implement SAC Algorithm
• Implement Temperature update
• Implement IMPALA Architecture