odd config

Author: jsuarez5341

config/boxoban.ini

@@ -3,45 +3,42 @@ env_name = boxoban
 
 [vec]
 total_agents = 16384
-num_buffers = 8
-num_threads = 8 
-
-[policy]
-num_layers = 1 
-hidden_size = 256
+num_buffers = 1.83193
+num_threads = 1
 
 [env]
 num_agents = 1
-#0 basic, 1 easy, 2 medium, 3 hard, 4 unfiltered
-difficulty = 1
-#reward per intermediate target (once per episode)
+difficulty = 2
 int_r_coeff = 0.25
-#moving box off target
-target_loss_pen_coeff = 0.0
+target_loss_pen_coeff = 0
 max_steps = 300
- 
+
+[policy]
+hidden_size = 1024
+num_layers = 7.61945
+expansion_factor = 1
+
 [train]
-anneal_lr = 1
-beta1 = 0.9774372816193448
-beta2 = 0.9659403664380584
-clip_coef = 0.6046560670053024
-ent_coef = 0.00002079831529141607
-eps = 0.00000000000001
-gae_lambda = 0.9258914518467392
-gamma = 0.9772998708784648
 gpus = 1
-horizon = 64
-learning_rate = 0.004480255741933225
-max_grad_norm = 1.221684008665154
-min_lr_ratio = 0.37872027027338984
-minibatch_size = 8192
-prio_alpha = 1
-prio_beta0 = 0.8789921736378042
-replay_ratio = 3.210300031048168
 seed = 42
-total_timesteps = 55504884
-use_rnn = true
-vf_clip_coef = 4.339748010438874
-vf_coef = 4.240274862679744
-vtrace_c_clip = 1.3625779006162615
-vtrace_rho_clip = 3.17260199042977
+total_timesteps = 1310719
+learning_rate = 0.0453785
+anneal_lr = 1
+min_lr_ratio = 0.37872
+gamma = 0.991084
+gae_lambda = 0.2
+replay_ratio = 4
+clip_coef = 0.01
+vf_coef = 3.98419
+vf_clip_coef = 5
+max_grad_norm = 0.1
+ent_coef = 0.000305154
+beta1 = 0.986866
+beta2 = 0.9
+eps = 1.03398e-14
+minibatch_size = 16384
+horizon = 32
+vtrace_rho_clip = 2.06753
+vtrace_c_clip = 2.50303
+prio_alpha = 0.907571
+prio_beta0 = 0.989874