vision test

pufferlib/config/ocean/vision_test.ini

@@ -0,0 +1,32 @@
+[base]
+package = ocean
+env_name = puffer_vision_test
+policy_name = Policy
+rnn_name = Recurrent
+
+[env]
+num_envs = 4096
+
+[train]
+total_timesteps = 150_000_000
+adam_beta1 = 0.9051396950168318
+adam_beta2 = 0.9998075206648028
+adam_eps = 6.080590274937272e-8
+anneal_lr = 1
+batch_size = auto
+bptt_horizon = 64
+clip_coef = 0.18136589604817419
+ent_coef = 0.008771869770071427
+gae_lambda = 0.852233401446272
+gamma = 0.9937244196032619
+learning_rate = 0.013637100389343416
+max_grad_norm = 3.8161147569794194 
+max_minibatch_size = 32768
+minibatch_size = 32768
+optimizer = muon
+prio_alpha = 0.6199253864227026
+prio_beta0 = 0.835222519905636
+vf_clip_coef = 0.1
+vf_coef = 1.304631362812606
+vtrace_c_clip = 1.4932276946661736
+vtrace_rho_clip = 0.45768721432684095

pufferlib/ocean/environment.py

@@ -157,6 +157,7 @@ def make_multiagent(buf=None, **kwargs):
     'asteroids': 'Asteroids',
     'whisker_racer': 'WhiskerRacer',
     'spaces': make_spaces,
+    'vision_test': 'VisionTest',
     'multiagent': make_multiagent,
 }
 

pufferlib/ocean/vision_test/binding.c

@@ -0,0 +1,15 @@
+#include "vision_test.h"
+
+#define Env VisionTest
+#include "../env_binding.h"
+
+static int my_init(Env* env, PyObject* args, PyObject* kwargs) {
+    return 0;
+}
+
+static int my_log(PyObject* dict, Log* log) {
+    assign_to_dict(dict, "episode_length", log->episode_length);
+    assign_to_dict(dict, "score", log->score);
+    assign_to_dict(dict, "n", log->n);
+    return 0;
+}

pufferlib/ocean/vision_test/vision_test.c

@@ -0,0 +1,36 @@
+#include "raylib.h"
+#include "rcamera.h"
+#include "raymath.h"
+#include "vision_test.h"
+#include "puffernet.h"
+
+int main(void) {
+    VisionTest env = {};
+
+    env.observations = (uint8_t*)calloc(3*64 + 1, sizeof(uint8_t));  // Alloc our 16x16 window compressed to float
+    env.actions = (int*)calloc(5, sizeof(int));
+    env.rewards = (float*)calloc(1, sizeof(float));
+    env.terminals = (unsigned char*)calloc(1, sizeof(unsigned char));
+
+    c_reset(&env);
+    c_render(&env);
+    while (!WindowShouldClose()) {
+        if (IsKeyDown(KEY_LEFT_SHIFT)) {
+            env.actions[0] = 0;
+            if      (IsKeyDown(KEY_W))    { env.actions[0] = S_UP;    }
+            else if (IsKeyDown(KEY_S))  { env.actions[0] = S_DOWN;  }
+            else if (IsKeyDown(KEY_A))  { env.actions[0] = S_LEFT;  }
+            else if (IsKeyDown(KEY_D)) { env.actions[0] = S_RIGHT; }
+        } else {
+            env.actions[0] = rand() % 5;
+        }
+
+        c_step(&env);
+        c_render(&env);
+    }
+    free(env.observations);
+    free(env.actions);
+    free(env.rewards);
+    free(env.terminals);
+    c_close(&env);
+}