Env - Artillery3D

pufferlib/config/ocean/artillery3d.ini

@@ -0,0 +1,184 @@
+[base]
+package = ocean
+env_name = puffer_artillery3d
+policy_name = Policy
+rnn_name = Recurrent
+
+[vec]
+num_envs = 8
+
+[env]
+num_envs = 1024
+debug = 0
+
+dist_fade = 0.75
+max_dist0 = 240
+max_reward = 1.00
+max_reward_dist = 5
+miss_penalty = -0.75
+out_bounds_penalty = -0.01
+target_size = 15
+turn_penalty = -0.003
+turn_penalty_delay = 100
+turn_penalty_ramp = 0.01
+rng = 7
+same_runs = 1
+
+[policy]
+hidden_size = 128
+
+[rnn]
+input_size = 128
+hidden_size = 128
+
+[train]
+adam_beta1 = 0.9603485277769915
+adam_beta2 = 0.9676992387435311
+adam_eps = 9.4852806860935e-08
+batch_size = auto
+bptt_horizon = 64
+clip_coef = 0.048815198297704664
+ent_coef = 0.049999999999999996
+gae_lambda = 0.95
+gamma = 0.9840496436204362
+learning_rate = 0.005
+max_grad_norm = 2.589981415830053
+#max_minibatch_size = 32768
+#min_minibatch_size = 32768
+minibatch_size = 16384 #16384 32768
+prio_alpha = 0.8573365561813164
+prio_beta0 = 0.9145256572446854
+total_timesteps = 400_000_000
+vf_clip_coef = 0.1921803059597499
+vf_coef = 0.8111721600187421
+vtrace_c_clip = 2.13335526866267
+vtrace_rho_clip = 2.0368704261583224
+
+[sweep]
+method = Protein
+metric = acc1000
+goal = maximize
+downsample = 10
+
+#[sweep.train.total_timesteps]
+#distribution = log_normal
+#min = 90_000_000
+#max = 100_000_000
+#mean = 90_500_000
+#scale = auto
+
+#[sweep.train.learning_rate]
+#distribution = log_normal
+#min = 0.005
+#mean = 0.0065
+#max = 0.01
+#scale = 0.5
+
+#[sweep.env.frameskip]
+#distribution = categorical
+#values = 1, 2, 4, 8
+
+[sweep.env.max_dist0]
+distribution = uniform
+min = 25.0
+max = 250.0
+mean = 200.0
+scale = auto
+
+[sweep.env.dist_fade]
+distribution = uniform
+min = 0.01
+max = 1.0
+mean = 0.36
+scale = auto
+
+#[sweep.env.same_runs]
+#distribution = uniform
+#min =1
+#max = 50
+#mean = 10
+#scale = auto
+
+[sweep.env.max_reward_dist]
+distribution = uniform
+min = 5
+max = 50
+mean = 10
+scale = auto
+
+[sweep.env.turn_penalty]
+distribution = uniform
+min = -0.01
+max = -0.003
+mean = -0.003
+scale = auto
+
+[sweep.env.turn_penalty_delay]
+distribution = uniform
+min = 50
+max = 100
+mean = 64
+scale = auto
+
+[sweep.env.turn_penalty_ramp]
+distribution = uniform
+min = 0.01
+max = 0.05
+mean = 0.023
+scale = auto
+
+[sweep.env.miss_penalty]
+distribution = uniform
+min = -1.0
+max = -0.1
+mean = -0.1
+scale = auto
+
+#[sweep.train.vtrace_rho_clip]
+#distribution = uniform
+#min = 1.0
+#max = 5.0
+#mean = 1.5
+#scale = auto
+
+#[sweep.train.max_grad_norm]
+#distribution = uniform
+#min = 1.0
+#mean = 1.15
+#max = 5.0
+#scale = auto
+
+#[sweep.train.vf_coef]
+#distribution = uniform
+#min = 0.0
+#max = 1.0
+#mean = 1.0
+#scale = auto
+
+#[sweep.train.gae_lambda]
+#distribution = logit_normal
+#min = 0.95
+#mean = 0.96
+#max = 0.9995
+#scale = auto
+
+#[sweep.train.vtrace_c_clip]
+#distribution = uniform
+#min = 0.0
+#max = 4.0
+#mean = 1.3
+#scale = auto
+
+#[sweep.train.minibatch_size]
+#distribution = uniform_pow2
+#min = 8192
+#max = 32768
+#mean = 16384
+#scale = auto
+
+#[sweep.train.ent_coef]
+#distribution = log_normal
+#min = 0.00001
+#mean = 0.018
+#max = 0.05
+#scale = auto

pufferlib/ocean/artillery3d/artillery3d.c

@@ -0,0 +1,60 @@
+#include "artillery3d.h"
+#include "puffernet.h"
+
+void demo() {
+    Artillery3D env = {
+        .debug = 0,
+        .dist_fade = 0.15,
+        .miss_penalty = -0.1,
+        .max_reward = 1.0,
+        .max_reward_dist = 5,
+        .max_dist0 = 250,
+        .out_bounds_penalty = -0.01,
+        .target_size = 15,
+        .turn_penalty = -0.003,
+        .turn_penalty_delay = 64,
+        .turn_penalty_ramp = 0.01,
+        .render = 1,
+        .rng = 7,
+        .same_runs = 1,
+        .i = 1,
+    };
+    allocate(&env);
+
+    env.client = make_client(&env);
+
+    const char* weights_path = "resources/artillery3d/puffer_artillery3d_weights.bin";
+    int weights_size = 135688;
+
+    Weights* weights = load_weights(weights_path, weights_size);
+    int logit_sizes[1] = {7};
+    int obs_size = 19;
+    LinearLSTM* net = make_linearlstm(weights, 1, obs_size, logit_sizes, 1);
+
+    c_reset(&env);
+    SetTargetFPS(30);
+    while (!WindowShouldClose()) {
+        if (IsKeyDown(KEY_LEFT_SHIFT)) {
+            env.actions[0] = 100;
+            if (IsKeyDown(KEY_SPACE)) env.actions[0] = 0;
+            if (IsKeyDown(KEY_UP)  || IsKeyDown(KEY_W)) env.actions[0] = 1;
+            if (IsKeyDown(KEY_DOWN)  || IsKeyDown(KEY_S)) env.actions[0] = 2;
+            if (IsKeyDown(KEY_LEFT)  || IsKeyDown(KEY_A)) env.actions[0] = 3;
+            if (IsKeyDown(KEY_RIGHT) || IsKeyDown(KEY_D)) env.actions[0] = 4;
+        } else {
+            int* actions = (int*)env.actions;
+            forward_linearlstm(net, env.observations, actions);
+            env.actions[0] = actions[0];
+        }
+        c_step(&env);
+        c_render(&env);
+    }
+    free_linearlstm(net);
+    free(weights);
+    free_allocated(&env);
+    close_client(env.client);
+}
+
+int main() {
+    demo();
+}