all training complete

a2c.py

@@ -259,7 +259,7 @@ def _thunk():
         episode_rewards = np.zeros((nenvs, ))
         final_rewards   = np.zeros((nenvs, ))
 
-        for update in tqdm(range(load_count + 1, total_timesteps)):
+        for update in tqdm(range(load_count + 1, total_timesteps + 1)):
             # mb stands for mini batch
             mb_obs, mb_rewards, mb_actions, mb_values, mb_dones = [],[],[],[],[]
             for n in range(nsteps):
@@ -331,10 +331,11 @@ def _thunk():
 if __name__ == '__main__':
     os.environ["CUDA_VISIBLE_DEVICES"]="0"
 
-    load_count = 100000
-    load_path = 'weights/model_%i.ckpt' % load_count
+    load_count = 0
+    load_path = 'weights/a2c_%i.ckpt' % load_count
     load_path = None
 
-    train(CnnPolicy, 'a2c', load_count, load_path, './a2c_logs')
+    train(CnnPolicy, 'a2c', load_count=load_count, load_path=load_path,
+            log_path='./a2c_logs')
 
 

env_model.py

@@ -164,7 +164,7 @@ def __init__(self, imag_state, imag_reward, input_states, input_actions,
     os.environ["CUDA_VISIBLE_DEVICES"]="1"
     with tf.Session() as sess:
         actor_critic = get_actor_critic(sess, nenvs, nsteps, ob_space, ac_space, CnnPolicy, should_summary=False)
-        actor_critic.load('weights/model_100000.ckpt')
+        actor_critic.load('weights/a2c_200000.ckpt')
 
         with tf.variable_scope('env_model'):
             env_model = create_env_model(ob_space, num_actions, num_pixels, len(mode_rewards['regular']))

i2a.py

@@ -135,7 +135,7 @@ def get_cache_loaded_a2c(sess, nenvs, nsteps, ob_space, ac_space):
         with tf.variable_scope('actor'):
             g_actor_critic = get_actor_critic(sess, nenvs, nsteps, ob_space,
                     ac_space, CnnPolicy, should_summary=False)
-        g_actor_critic.load('weights/model_100000.ckpt')
+        g_actor_critic.load('weights/a2c_200000.ckpt')
         print('Actor restored!')
     return g_actor_critic