initialize inside function :(

main_noweb.py

@@ -33,7 +33,6 @@ os.system("nvidia-smi")
 print("[INFO]: Imported modules!")
 human = MMPoseInferencer("simcc_mobilenetv2_wo-deconv-8xb64-210e_coco-256x192") # simcc_mobilenetv2_wo-deconv-8xb64-210e_coco-256x192 dekr_hrnet-w32_8xb10-140e_coco-512x512
 hand = MMPoseInferencer("hand")
-human3d = MMPoseInferencer(pose3d="human3d")
 
 #"https://github.com/open-mmlab/mmpose/blob/main/configs/body_3d_keypoint/pose_lift/h36m/pose-lift_simplebaseline3d_8xb64-200e_h36m.py", 
                            #"https://download.openmmlab.com/mmpose/body3d/simple_baseline/simple3Dbaseline_h36m-f0ad73a4_20210419.pth") # pose3d="human3d"
@@ -47,7 +46,7 @@ def check_fps(video):
     return nframes, fps
 
 
-def get_frames(video, fps=15, height=512, width=512):
+def get_frames(video, fps=30, height=512, width=512):
 # extract the file name and extension
     split_tup = os.path.splitext(video)
 
@@ -61,14 +60,14 @@ def get_frames(video, fps=15, height=512, width=512):
     clip = moviepy.VideoFileClip(video)
     
     #check fps
-    if clip.fps > 15:
+    if clip.fps > fps:
         print("vide rate is over 15, resetting to 15")
         #clip_resized = moviepy.clip.resize(height=height)
         clip.write_videofile(video, fps=fps)
     else:
         print("video rate is OK")
         #clip_resized = clip.resize(height=height)
-        #clip_resized.write_videofile(video, fps=clip.fps)
+        clip.write_videofile(video, fps=clip.fps)
     
     print("video resized to 512 height")
   
@@ -93,7 +92,8 @@ def check_extension(video):
 def pose3d(video, kpt_threshold):
     video = check_extension(video)
     print(device)
-    temp_3d = human3d
+
+    human3d = MMPoseInferencer(pose3d="human3d")
 
     # Define new unique folder
     add_dir = str(uuid.uuid4())
@@ -101,9 +101,8 @@ def pose3d(video, kpt_threshold):
 
     os.makedirs(add_dir)
     print(check_fps(video))
-    result_generator = temp_3d(video, 
+    result_generator = human3d(video, 
                             vis_out_dir = add_dir,
-                            #return_vis=True,
                             radius = 5,
                             thickness=4,
                             rebase_keypoint_height=True,
@@ -117,7 +116,7 @@ def pose3d(video, kpt_threshold):
     out_file = glob.glob(os.path.join(add_dir, "*.mp4")) #+ glob.glob(os.path.join(vis_out_dir, "*.webm")) 
     kpoints = glob.glob(os.path.join(add_dir, "*.json"))
     print(kpoints) 
-    temp_3d = 0
+
     # Reinitialize
     return "".join(out_file)
 
@@ -185,19 +184,19 @@ def UI():
                 with gr.Column():
                     with gr.Row():
                         with gr.Column():
-                            video_input = gr.Video(source="upload", type="filepath", height=512, width=512)
-                            # Insert slider with kpt_thr
-                            file_kpthr = gr.Slider(0, 1, value=0.3, label='Keypoint threshold')
+                            with gr.Row():
+                                video_input = gr.Video(source="upload", type="filepath", height=512, width=512)
+                                # Insert slider with kpt_thr
+                                file_kpthr = gr.Slider(0, 1, value=0.3, label='Keypoint threshold')
                             with gr.Row():
                                 submit_pose_file = gr.Button("Make 2d pose estimation")
                                 submit_pose3d_file = gr.Button("Make 3d pose estimation")
                                 submit_hand_file = gr.Button("Make 2d hand estimation")
-                                submit_detect_file = gr.Button("Detect and track objects")
 
                     with gr.Row():
-                        video_output1 = gr.PlayableVideo(height=512,  label = "Estimate human 2d poses", show_label=True)
+                        video_output1 = gr.Video(height=512,  label = "Estimate human 2d poses", show_label=True)
                         video_output2 = gr.Video(height=512,  label = "Estimate human 3d poses", show_label=True)
-                        video_output3 = gr.PlayableVideo(height=512,  label = "Estimate human hand poses", show_label=True)
+                        video_output3 = gr.Video(height=512,  label = "Estimate human hand poses", show_label=True)
                     
                     gr.Markdown("Download the .json file that contains the keypoint positions for each frame in the video.")
                     jsonoutput = gr.File(file_types=[".json"])