updates

rogeriobonatti · Jun 5, 2022 · 4323728 · 4323728
1 parent 198fda6
commit 4323728
Show file tree

Hide file tree

Showing 4 changed files with 71 additions and 38 deletions.
diff --git a/mushr_rhc_ros/launch/map_server_test_env.launch b/mushr_rhc_ros/launch/map_server_test_env.launch
@@ -0,0 +1,19 @@
+<launch>
+    <!-- <arg name="map" default="$(find mushr_rhc_ros)/maps/real-floor0-edited.yaml" /> -->
+    <!-- <arg name="c2g_map" default="$(find mushr_rhc_ros)/maps/real-floor0-edited-c2g-full-loop.yaml" /> -->
+
+    <!-- <arg name="map" default="$(find mushr_rhc_ros)/maps/bravern_13_mod.yaml" />
+    <arg name="c2g_map" default="$(find mushr_rhc_ros)/maps/bravern_13_mod.yaml" /> -->
+
+    <arg name="map" default="$(find mushr_rhc_ros)/maps/map_12f_cropped.yaml" />
+    <arg name="c2g_map" default="$(find mushr_rhc_ros)/maps/map_12f_cropped.yaml" />
+
+    <node pkg="map_server" name="map_server" type="map_server" args="$(arg map)"/>
+    <node pkg="map_server" name="c2g_map" type="map_server" args="$(arg c2g_map)">
+        <remap from="/static_map" to="/c2g/static_map" />
+        <remap from="/map_metadata" to="/c2g/map_metadata" />
+        <remap from="/map" to="/c2g/map" />
+    </node>
+
+    <param name="map_file" value="$(arg c2g_map)" />
+</launch>
diff --git a/mushr_rhc_ros/launch/sim/sim_server_eval.launch b/mushr_rhc_ros/launch/sim/sim_server_eval.launch
@@ -3,38 +3,44 @@
 	<arg name="map_server" default="1" />
 	<arg name="car_name" default="car" />
 
-    <!-- <arg name="out_path" default="/home/rb/hackathon_data/e2e_eval/model_test" />
-    <arg name="model_path" default="/home/rb/hackathon_data/aml_outputs/log_output/normal-kingfish/GPTiros_e2e_8gpu_2022-02-17_1645120431.7528405_2022-02-17_1645120431.7528613/model/epoch10.pth.tar" /> -->
+    <!-- logging output paths -->
+    <arg name="out_path" default="/home/rb/hackathon_data_premium/e2e_eval/model_test" />
 
-    <!-- current model! -->
-    <!-- <arg name="out_path" default="/home/rb/hackathon_data/e2e_eval/model_test" />
-    <arg name="model_path" default="/home/rb/hackathon_data/aml_outputs/log_output/deep-adder/GPTnips_e2e_2022-04-06_1649220643.529554_2022-04-06_1649220643.5295672/model/epoch21.pth.tar" /> -->
+    <!-- action model -->
+    <arg name="model_path_act" default="/home/rb/hackathon_data_premium/aml_outputs/log_output/hvd_test_16/GPTcorl_scratch_trainm_e2e_statet_pointnet_traini_1_nla_12_nhe_8_statel_0.01_2022-06-02_1654131996.2524076_2022-06-02_1654131996.2524228/model/epoch30.pth.tar" />
 
-    <!-- <arg name="out_path" default="/home/rb/hackathon_data/e2e_eval/model_test" />
-    <arg name="model_path" default="/home/rb/hackathon_data/aml_outputs/log_output/renewed-colt/GPTnips_8gpu_relu_e2e_slow_bias_2022-04-02_1648869033.3693378_2022-04-02_1648869033.36935/model/epoch22.pth.tar" /> -->
+    <!-- map model -->
+    <arg name="use_map" default="true" />
+    <arg name="model_path_map" default="/home/rb/hackathon_data_premium/aml_outputs/log_output/mapscratch_new_0/GPTcorl_map_trainm_map_sta_pointnet_traini_1_nla_12_nhe_8_2022-05-31_1653978768.732001_2022-05-31_1653978768.7320147/model/epoch28.pth.tar" />
 
-    <!-- <arg name="out_path" default="/home/rb/hackathon_data/e2e_eval/model_test" />
-    <arg name="model_path" default="/home/rb/hackathon_data/aml_outputs/log_output/mushr_nips/GPTnips_8gpu_relu_e2e_slow_2022-03-31_1648698385.6160257_2022-03-31_1648698385.6160388/model/epoch5.pth.tar" /> -->
+    <!-- localization model -->
+    <arg name="use_loc" default="false" />
+    <arg name="model_path_loc" default="/home/rb/hackathon_data_premium/aml_outputs/log_output/locscratch_new_0/GPTcorl_loc_trainm_loc_sta_pointnet_lr_6e-5_traini_1_nla_12_nhe_8_locx_0.01_locy_1_loca_10_locd_joint_2022-05-31_1653978601.5423563_2022-05-31_1653978601.5423756/model/epoch30.pth.tar" />
 
-    <!-- current model! -->
-    <!-- <arg name="out_path" default="/home/rb/hackathon_data/e2e_eval/model_test" />
-    <arg name="model_path" default="/home/rb/hackathon_data/aml_outputs/log_output/tough-mongoose/GPTnips_e2e_2022-04-07_1649355173.7379868_2022-04-07_1649355173.7379994/model/epoch22.pth.tar" /> -->
+    <arg name="deployment_map" default="train" />
+    <!-- <arg name="deployment_map" default="test" /> -->
 
-    <!-- current model! -->
-    <arg name="out_path" default="/home/rb/hackathon_data_premium/e2e_eval/model_test" />
-    <arg name="model_path" default="/home/rb/hackathon_data_premium/aml_outputs/log_output/pretrain_new_0/GPTcorl_scratch_trainm_e2e_statet_pointnet_traini_1_nla_12_nhe_8_statel_0.01_2022-05-31_1654007043.7606997_2022-05-31_1654007043.760718/model/epoch14.pth.tar" />
-    <!-- <arg name="model_path" default="/home/rb/hackathon_data_premium/aml_outputs/log_output/pretrain_scratch_0/GPTcorl_scratch_trainm_e2e_statet_pointnet_traini_1_nla_6_nhe_8_statel_0.01_2022-05-23_1653348711.3057182_2022-05-23_1653348711.3057299/model/epoch12.pth.tar" /> -->
+    <group if="$(eval arg('deployment_map') == 'train')">
+	    <include file="$(find mushr_rhc_ros)/launch/map_server.launch" />
+    </group>
 
-    <group if="$(arg map_server)">
-	<include file="$(find mushr_rhc_ros)/launch/map_server.launch" />
+    <group if="$(eval arg('deployment_map') == 'test')">
+	    <include file="$(find mushr_rhc_ros)/launch/map_server_test_env.launch" />
     </group>
 
     <group ns="$(arg car_name)">
         <node pkg="mushr_rhc_ros" type="rhcnode_network_pcl_new.py" name="rhcontroller" output="screen">
             <env name="RHC_USE_CUDA" value="0" />
 
+            <param name="deployment_map" value="$(arg deployment_map)" />
+
             <param name="out_path" value="$(arg out_path)" />
-            <param name="model_path" value="$(arg model_path)" />
+            <param name="model_path_act" value="$(arg model_path_act)" />
+            <param name="model_path_map" value="$(arg model_path_map)" />
+            <param name="model_path_loc" value="$(arg model_path_loc)" />
+
+            <param name="use_map" value="$(arg use_map)" />
+            <param name="use_loc" value="$(arg use_loc)" />
 
             <param name="inferred_pose_t" value="car_pose" />
 

diff --git a/mushr_rhc_ros/src/rhcnode_network_pcl.py b/mushr_rhc_ros/src/rhcnode_network_pcl.py
@@ -152,7 +152,7 @@ def __init__(self, dtype, params, logger, name):
         # mapping model
         if self.use_map:
 
-            saved_map_model_path = '/home/rb/hackathon_data_premium/aml_outputs/log_output/mapscratch_new_0/GPTcorl_map_trainm_map_sta_pointnet_traini_1_nla_12_nhe_8_2022-05-31_1653978768.732001_2022-05-31_1653978768.7320147/model/epoch8.pth.tar'
+            saved_map_model_path = '/home/rb/hackathon_data_premium/aml_outputs/log_output/mapscratch_new_0/GPTcorl_map_trainm_map_sta_pointnet_traini_1_nla_12_nhe_8_2022-05-31_1653978768.732001_2022-05-31_1653978768.7320147/model/epoch28.pth.tar'
 
             mconf_map = GPTConfig(block_size, max_timestep,
                       n_layer=12, n_head=8, n_embd=128, model_type='GPT', use_pred_state=True,
@@ -182,7 +182,7 @@ def __init__(self, dtype, params, logger, name):
         # localization model
         if self.use_loc:
 
-            saved_loc_model_path = '/home/rb/hackathon_data_premium/aml_outputs/log_output/locscratch_new_0/GPTcorl_loc_trainm_loc_sta_pointnet_lr_6e-5_traini_1_nla_12_nhe_8_locx_0.01_locy_1_loca_10_locd_joint_2022-05-31_1653978601.5423563_2022-05-31_1653978601.5423756/model/epoch8.pth.tar'
+            saved_loc_model_path = '/home/rb/hackathon_data_premium/aml_outputs/log_output/locscratch_new_0/GPTcorl_loc_trainm_loc_sta_pointnet_lr_6e-5_traini_1_nla_12_nhe_8_locx_0.01_locy_1_loca_10_locd_joint_2022-05-31_1653978601.5423563_2022-05-31_1653978601.5423756/model/epoch26.pth.tar'
 
             mconf_loc = GPTConfig(block_size, max_timestep,
                       n_layer=12, n_head=8, n_embd=128, model_type='GPT', use_pred_state=True,
@@ -588,7 +588,7 @@ def prepare_model_inputs(self):
 
         self.act_lock.acquire()
         for act in self.q_actions.queue:
-            x_act[0,idx] = torch.tensor(act)
+            x_act[0,idx] = torch.tensor(pre.norm_angle(act))
             idx+=1
         self.act_lock.release()
 

diff --git a/mushr_rhc_ros/src/rhcnode_network_pcl_new.py b/mushr_rhc_ros/src/rhcnode_network_pcl_new.py
@@ -84,8 +84,7 @@ def __init__(self, dtype, params, logger, name):
         self.default_angle = 0.0
         self.nx = None
         self.ny = None
-        self.use_map = False
-        self.use_loc = True
+
         self.points_viz_list = None
         self.map_recon = None
         self.loc_counter = 0
@@ -101,8 +100,12 @@ def __init__(self, dtype, params, logger, name):
         self.device = device
         self.clip_len = 16
 
-        # tests for IROS
-        saved_model_path = rospy.get_param("~model_path", 'default_value')
+        self.map_type = rospy.get_param("~deployment_map", 'train')
+
+        self.use_map = rospy.get_param("~use_map", False)
+        self.use_loc = rospy.get_param("~use_loc", False)
+
+        saved_model_path_action = rospy.get_param("~model_path_act", 'default_value')
         self.out_path = rospy.get_param("~out_path", 'default_value')
 
         vocab_size = 100
@@ -116,11 +119,11 @@ def __init__(self, dtype, params, logger, name):
                       map_decoder='deconv', map_recon_dim=64, freeze_core=False,
                       state_loss_weight=0.1,
                       loc_x_loss_weight=0.01, loc_y_loss_weight=0.1, loc_angle_loss_weight=10.0,
-                      loc_decoder_type='separate')
+                      loc_decoder_type='joint')
         model = GPT(mconf, device)
         # model=torch.nn.DataParallel(model)
 
-        checkpoint = torch.load(saved_model_path, map_location=device)
+        checkpoint = torch.load(saved_model_path_action, map_location=device)
         # old code for loading model
         model.load_state_dict(checkpoint['state_dict'])
         # new code for loading mode
@@ -153,8 +156,8 @@ def __init__(self, dtype, params, logger, name):
 
         # mapping model
         if self.use_map:
-
-            saved_map_model_path = '/home/rb/hackathon_data_premium/aml_outputs/log_output/mapscratch_new_0/GPTcorl_map_trainm_map_sta_pointnet_traini_1_nla_12_nhe_8_2022-05-31_1653978768.732001_2022-05-31_1653978768.7320147/model/epoch8.pth.tar'
+            
+            saved_map_model_path = rospy.get_param("~model_path_map", '')
 
             mconf_map = GPTConfig(block_size, max_timestep,
                       n_layer=12, n_head=8, n_embd=128, model_type='GPT', use_pred_state=True,
@@ -183,8 +186,8 @@ def __init__(self, dtype, params, logger, name):
 
         # localization model
         if self.use_loc:
-
-            saved_loc_model_path = '/home/rb/hackathon_data_premium/aml_outputs/log_output/locscratch_new_0/GPTcorl_loc_trainm_loc_sta_pointnet_lr_6e-5_traini_1_nla_12_nhe_8_locx_0.01_locy_1_loca_10_locd_joint_2022-05-31_1653978601.5423563_2022-05-31_1653978601.5423756/model/epoch17.pth.tar'
+            
+            saved_loc_model_path = rospy.get_param("~model_path_loc", '')
 
             mconf_loc = GPTConfig(block_size, max_timestep,
                       n_layer=12, n_head=8, n_embd=128, model_type='GPT', use_pred_state=True,
@@ -232,7 +235,7 @@ def __init__(self, dtype, params, logger, name):
         # set timer callbacks for visualization
         rate_map_display = 1.0
         rate_loc_display = 20
-        # self.map_viz_timer = rospy.Timer(rospy.Duration(1.0 / rate_map_display), self.map_viz_cb)
+        self.map_viz_timer = rospy.Timer(rospy.Duration(1.0 / rate_map_display), self.map_viz_cb)
         self.map_viz_loc = rospy.Timer(rospy.Duration(1.0 / rate_loc_display), self.loc_viz_cb)
 
 
@@ -249,8 +252,11 @@ def start(self):
         self.logger.info("Initialized")
 
         # set initial pose for the car in the very first time in an allowable region
-        self.send_initial_pose()
-        # self.send_initial_pose_12f()
+        if self.map_type == 'train':
+                self.send_initial_pose()
+        else:
+            self.send_initial_pose_12f()
+
         self.time_started = rospy.Time.now()
 
         # wait until we actually have a car pose
@@ -544,7 +550,7 @@ def prepare_model_inputs(self, queue_type):
             self.large_queue_lock.release()
         for idx, element in enumerate(queue_list):
             x_imgs[0,idx,:] = torch.tensor(element[0])
-            x_act[0,idx] = torch.tensor(element[1])
+            x_act[0,idx] = torch.tensor(pre.norm_angle(element[1]))
         # x_imgs = x_imgs.contiguous().view(1, self.clip_len, 200*200)
         x_imgs = x_imgs.to(self.device)
         x_act = x_act.view(1, self.clip_len , 1)
@@ -579,8 +585,10 @@ def check_reset(self, rate_hz):
             print("Distance: {}  | Time: {} | Time so far: {}".format(self.distance_so_far, delta_time, self.time_so_far))
             with open(self.file_name,'a') as fd:
                 fd.write(str(self.distance_so_far)+','+str(self.time_so_far)+'\n')
-            self.send_initial_pose()
-            # self.send_initial_pose_12f()
+            if self.map_type == 'train':
+                self.send_initial_pose()
+            else:
+                self.send_initial_pose_12f()
             rospy.loginfo("Got stuck, resetting pose of the car to default value")
             msg = String()
             msg.data = "got stuck"