From 60d51667e2fd0f460bbcbefcaeddecdecac94ca1 Mon Sep 17 00:00:00 2001
From: likejiao <likejiao@baidu.com>
Date: Mon, 16 Mar 2020 19:53:44 +0800
Subject: [PATCH] update readme

---
 benchmark/torch/coma/README.md | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/benchmark/torch/coma/README.md b/benchmark/torch/coma/README.md
index 75d509b..cfa3562 100644
--- a/benchmark/torch/coma/README.md
+++ b/benchmark/torch/coma/README.md
@@ -3,15 +3,16 @@
 This is an PARL + PyTorch implementation of the multi-agent reinforcement learning algorithms: COMA.
 
 ### Paper
-
 - [Counterfactual Multi-Agent Policy Gradients](https://arxiv.org/abs/1705.08926)
 
-
 ### Benchmark Result
 Mean win_rate (evaluate 5 episode) for 1000 epchos training (1 epcho = 5 episodes).
 
 <img src=".benchmark/3m_result.png" width = "400" height = "300" alt="coma-3m"/>
 
+
+
+
 ## StarCraft II Installation
 The environment is based on the full game of StarCraft II (version >= 3.16.1). To install the game, follow the commands bellow, or check more detail in [SMAC](https://github.com/oxwhirl/smac#installing-starcraft-ii)
 
@@ -27,9 +28,10 @@ $ bash build_docker.sh  # build the Dockerfile
 $ bash install_sc2.sh  # download startcraft II and maps
 ```
 
+
+
 ## How to use
 ### Dependencies
-
 - python3.5+
 - parl
 - torch
@@ -46,3 +48,9 @@ $ cd coma
 $ NV_GPU=$your_gpu_id docker run --name $your_container_name --user $(id -u):$(id -g) -v `pwd`:/parl -t parl-starcraft2:1.0 python3 train.py
 ```
 *or you can operate docker interactively by `docker run --name $your_container_name -it -v $your_host_path:/parl -t parl-starcraft2:1.0  /bin/bash`*
+
+
+
+### Reference
+- [StarCraft](https://github.com/starry-sky6688/StarCraft)
+- [pymarl](https://github.com/oxwhirl/pymarl)
-- 
GitLab