battle env

Author: Joseph Suarez

pufferlib/config/ocean/battle.ini

@@ -12,11 +12,11 @@ input_size = 512
 hidden_size = 512
 
 [vec]
-num_envs = 16
+num_envs = 8
 
 [env]
-num_envs = 2
-num_agents = 512
+num_envs = 8
+num_agents = 128
 num_armies = 2
 size_x = 2
 size_y = 2
@@ -27,7 +27,7 @@ total_timesteps = 100_000_000
 
 learning_rate = 0.0015534438005054883
 gamma = 0.9923382806478448
-minibatch_size = 65536
+minibatch_size = 32768
 
 #adam_beta1 = 0.5797997352318079
 #adam_beta2 = 0.9001752474216785
@@ -45,3 +45,5 @@ minibatch_size = 65536
 #vf_coef = 1.8024088377114245
 #vtrace_c_clip = 1.7578256946375268
 #vtrace_rho_clip = 1.0041987439042879
+
+

pufferlib/ocean/battle/battle.c

@@ -31,7 +31,7 @@ int main() {
     init(&env);
 
     // Allocate these manually since they aren't being passed from Python
-    int num_obs = 6*env.num_armies + 19 + 8;
+    int num_obs = 3*env.num_armies + 4*16 + 22 + 8;
     env.observations = calloc(env.num_agents*num_obs, sizeof(float));
     env.actions = calloc(3*env.num_agents, sizeof(int));
     env.rewards = calloc(env.num_agents, sizeof(float));
@@ -44,6 +44,7 @@ int main() {
     int ctrl = 0;
 
     while (!WindowShouldClose()) {
+        /*
         for (int i=0; i<env.num_agents; i++) {
             Entity* agent = &env.agents[i];
             int army = agent->army;
@@ -104,6 +105,7 @@ int main() {
             //env.actions[3*i + 1] = 4.0f;
             //env.actions[3*i + 2] = 4.0f;
         }
+        */
 
         if (IsKeyDown(KEY_LEFT_SHIFT)) {
             if (IsKeyPressed(KEY_TAB)) {