creating pose from video

handler.py

@@ -273,7 +273,7 @@ class EndpointHandler():
         video_url = inputs.get("video_url", "")
         width = inputs.get("width", 512)
         height = inputs.get("height", 768)
-        length = inputs.get("length", 96)
+        length = inputs.get("length", 144)
         num_inference_steps = inputs.get("num_inference_steps", 25)
         cfg = inputs.get("cfg", 3.5)
         seed = inputs.get("seed", 123)
@@ -340,9 +340,9 @@ class EndpointHandler():
 
             torch.manual_seed(seed)
 
-            temporary_pose_video_path = os.path.join(base_dir, 'dancing.mp4')
-            pose_images = read_frames(temporary_pose_video_path)
-            src_fps = get_fps(temporary_pose_video_path)
+            # temporary_pose_video_path = os.path.join(base_dir, 'dancing.mp4')
+            pose_images = read_frames(downloaded_video_path)
+            src_fps = get_fps(downloaded_video_path)
             
             pose_list = []
             total_length = min(length, len(pose_images))
@@ -394,13 +394,13 @@ class EndpointHandler():
 
             #remove background
             # self.print_directory_contents()
-            removed_background_output_path = os.path.join(save_dir, "removed_background_result.mp4")
-            remove_background_script_path = os.path.join(base_dir, "rembg_video.py")
-            remove_background_command = f'python3 {remove_background_script_path} {swapped_face_video_path} {removed_background_output_path}'
-            print("Command is " + remove_background_command)
-            remove_background_result = subprocess.run(remove_background_command, shell=True, capture_output=True, text=True)
-            if remove_background_result.returncode != 0:
-                raise RuntimeError(f"Error running removing backgriund: {remove_background_result.stderr}")
+            # removed_background_output_path = os.path.join(save_dir, "removed_background_result.mp4")
+            # remove_background_script_path = os.path.join(base_dir, "rembg_video.py")
+            # remove_background_command = f'python3 {remove_background_script_path} {swapped_face_video_path} {removed_background_output_path}'
+            # print("Command is " + remove_background_command)
+            # remove_background_result = subprocess.run(remove_background_command, shell=True, capture_output=True, text=True)
+            # if remove_background_result.returncode != 0:
+            #     raise RuntimeError(f"Error running removing backgriund: {remove_background_result.stderr}")
 
 
             # Perform RIFE interpolation
@@ -409,7 +409,7 @@ class EndpointHandler():
             # self.run_rife_interpolation(swapped_face_video_path, rife_output_path, multi=2, scale=0.5)
 
             # Encode the final video in base64
-            with open(removed_background_output_path, "rb") as video_file:
+            with open(swapped_face_video_path, "rb") as video_file:
                 video_base64 = base64.b64encode(video_file.read()).decode("utf-8")
 
             torch.cuda.empty_cache()

sampler.py

@@ -22,9 +22,9 @@ handler = EndpointHandler()
 # Define sample inputs
 inputs = {
     "inputs": {
-        "ref_image_url": "https://media.discordapp.net/attachments/1237667104866697267/1246982998520496261/image.jpg?ex=665e5eea&is=665d0d6a&hm=46a719438e039655788336902d62419a2921a00720b21ac17e74755cb6de8cee&=&format=webp&width=976&height=1192",
-        "video_url": "https://cdn.discordapp.com/attachments/1237667104866697267/1246950669689290752/pose.mp4?ex=665e40ce&is=665cef4e&hm=37a5f042a2d03f71c4ae5e9f56616a35d205b97df0ff205182cab28aa7784bbb&",
-        "length": 12,
+        "ref_image_url": "https://media.discordapp.net/attachments/1237667074210267217/1247311431028703242/image.png?ex=665f90cb&is=665e3f4b&hm=89767149275a6ef40e39d4a73d77019ab30ab665512398816bc614cfecdddb34&=&format=webp&quality=lossless&width=590&height=1194",
+        "video_url": "https://firebasestorage.googleapis.com/v0/b/quiz-app-edffe.appspot.com/o/creeper_shortneed_pose.mp4?alt=media&token=f87f6e8c-6ab2-445c-8b55-205f08f1f5d5",
+        "length": 96,
         "num_inference_steps": 25,
         "cfg": 3.5,
         "seed": 123