patch-3.7.1 (#1178)

* patch 3.7.1

* fix bug for 2 processors on image to image pipeline (#1177)

* reduce default value of target_frame_amount

* restore old performance

* restore old performance

* restore old performance

* restore old performance

facefusion/program.py

@@ -606,7 +606,7 @@ def create_frame_distribution_program() -> ArgumentParser:
 				'--target-frame-amount',
 				help = translator.get('help.target_frame_amount'),
 				type = int,
-				default = config.get_int_value('frame_distribution', 'target_frame_amount', '5'),
+				default = config.get_int_value('frame_distribution', 'target_frame_amount', '2'),
 				choices = facefusion.choices.target_frame_amount_range,
 				metavar = create_int_metavar(facefusion.choices.target_frame_amount_range)
 			)

facefusion/video_manager.py

@@ -19,6 +19,12 @@ def get_video_capture(video_path : str) -> cv2.VideoCapture:
 	return VIDEO_POOL_SET.get('capture').get(video_path)
 
 
+def conditional_set_video_frame_position(video_capture : cv2.VideoCapture, frame_position : int) -> bool:
+	if not video_capture.get(cv2.CAP_PROP_POS_FRAMES) == frame_position:
+		return video_capture.set(cv2.CAP_PROP_POS_FRAMES, frame_position)
+	return True
+
+
 def get_video_writer(video_path : str) -> cv2.VideoWriter:
 	if video_path not in VIDEO_POOL_SET.get('writer'):
 		video_writer = cv2.VideoWriter()

facefusion/vision.py

@@ -11,7 +11,7 @@ from facefusion.filesystem import get_file_extension, is_image, is_video
 from facefusion.media_helper import restrict_trim_frame
 from facefusion.thread_helper import thread_lock, thread_semaphore
 from facefusion.types import ColorMode, Duration, Fps, Mask, Orientation, Resolution, Scale, VisionFrame
-from facefusion.video_manager import get_video_capture
+from facefusion.video_manager import conditional_set_video_frame_position, get_video_capture
 
 
 def read_static_images(image_paths : List[str], color_mode : ColorMode = 'rgb') -> List[VisionFrame]:
@@ -81,11 +81,11 @@ def read_video_frame(video_path : str, frame_number : int = 0) -> Optional[Visio
 		video_capture = get_video_capture(video_path)
 
 		if video_capture and video_capture.isOpened():
-			frame_total = video_capture.get(cv2.CAP_PROP_FRAME_COUNT)
-			frame_position = min(frame_total, frame_number)
+			video_frame_total = int(video_capture.get(cv2.CAP_PROP_FRAME_COUNT))
+			video_frame_position = min(video_frame_total, frame_number)
 
 			with thread_semaphore():
-				video_capture.set(cv2.CAP_PROP_POS_FRAMES, frame_position)
+				conditional_set_video_frame_position(video_capture, video_frame_position)
 				has_vision_frame, vision_frame = video_capture.read()
 
 			if has_vision_frame:
@@ -106,13 +106,13 @@ def read_video_chunk(video_path : str, chunk_number : int, chunk_size : int) ->
 		video_capture = get_video_capture(video_path)
 
 		if video_capture and video_capture.isOpened():
-			frame_total = int(video_capture.get(cv2.CAP_PROP_FRAME_COUNT))
-			frame_position = chunk_number * chunk_size
+			video_frame_total = int(video_capture.get(cv2.CAP_PROP_FRAME_COUNT))
+			video_frame_position = chunk_number * chunk_size
 
 			with thread_semaphore():
-				video_capture.set(cv2.CAP_PROP_POS_FRAMES, frame_position)
+				conditional_set_video_frame_position(video_capture, video_frame_position)
 
-				for frame_number in range(frame_position, min(frame_position + chunk_size, frame_total)):
+				for frame_number in range(video_frame_position, min(video_frame_position + chunk_size, video_frame_total)):
 					has_vision_frame, vision_frame = video_capture.read()
 
 					if has_vision_frame:
@@ -121,9 +121,9 @@ def read_video_chunk(video_path : str, chunk_number : int, chunk_size : int) ->
 	return video_frame_chunk
 
 
-def select_video_frames(video_path : str, frame_number : int = 0, frame_offset : int = 5) -> List[VisionFrame]:
+def select_video_frames(video_path : str, frame_number : int = 0, frame_offset : int = 2) -> List[VisionFrame]:
 	vision_frames = []
-	chunk_size = (frame_offset * 2 + 1) * 4
+	chunk_size = frame_offset * 2 + 1
 
 	if is_video(video_path):
 		with thread_lock():

facefusion/workflows/core.py

@@ -77,10 +77,10 @@ def conditional_get_reference_vision_frame() -> VisionFrame:
 	return read_static_image(state_manager.get_item('target_path'))
 
 
-def conditional_get_target_vision_frames(temp_frame_path : str, frame_number : int) -> List[VisionFrame]:
+def conditional_get_target_vision_frames(frame_number : int) -> List[VisionFrame]:
 	if state_manager.get_item('workflow') in [ 'image-to-video', 'image-to-video:frames' ]:
 		return select_video_frames(state_manager.get_item('target_path'), frame_number, state_manager.get_item('target_frame_amount'))
-	return [ read_static_image(temp_frame_path) ]
+	return [ read_static_image(state_manager.get_item('target_path')) ]
 
 
 def process_temp_frame(temp_frame_path : str, frame_number : int) -> bool:
@@ -88,7 +88,7 @@ def process_temp_frame(temp_frame_path : str, frame_number : int) -> bool:
 	source_vision_frames = read_static_images(state_manager.get_item('source_paths'))
 	source_audio_frame = conditional_get_source_audio_frame(frame_number)
 	source_voice_frame = conditional_get_source_voice_frame(frame_number)
-	target_vision_frames = conditional_get_target_vision_frames(temp_frame_path, frame_number)
+	target_vision_frames = conditional_get_target_vision_frames(frame_number)
 	temp_vision_frame = read_static_image(temp_frame_path, 'rgba').copy()
 	temp_vision_mask = extract_vision_mask(temp_vision_frame)