optimization with mac

Hoping this would solve the mac issues, if you're a mac user, please report if there is an improvement

README.md

@@ -30,7 +30,7 @@ By using this software, you agree to these terms and commit to using it in a man
 
 Users are expected to use this software responsibly and legally. If using a real person's face, obtain their consent and clearly label any output as a deepfake when sharing online. We are not responsible for end-user actions.
 
-## Exclusive v2.2 Quick Start - Pre-built (Windows/Mac Silicon)
+## Exclusive v2.3c Quick Start - Pre-built (Windows/Mac Silicon)
 
   <a href="https://deeplivecam.net/index.php/quickstart"> <img src="media/Download.png" width="285" height="77" />
 
@@ -354,11 +354,13 @@ Looking for a CLI mode? Using the -s/--source argument will make the run program
  - [*"Alright look look look, now look chat, we can do any face we want to look like chat"*](https://www.youtube.com/live/mFsCe7AIxq8?feature=shared&t=2686) - IShowSpeed
  - [*"They do a pretty good job matching poses, expression and even the lighting"*](https://www.youtube.com/watch?v=wnCghLjqv3s&t=551s) - TechLinked (LTT)
  - [*"Als Sean Connery an der Redaktionskonferenz teilnahm"*](https://www.golem.de/news/deepfakes-als-sean-connery-an-der-redaktionskonferenz-teilnahm-2408-188172.html) - Golem.de (German)
+ - [*"What the F***! Why do I look like Vinny Jr? I look exactly like Vinny Jr!? No, this shit is crazy! Bro This is F*** Crazy! "*](https://youtu.be/JbUPRmXRUtE?t=3964) - IShowSpeed
 
 
 ## Credits
 
 -   [ffmpeg](https://ffmpeg.org/): for making video-related operations easy
+-   [Henry](https://github.com/henryruhs): One of the major contributor in this repo
 -   [deepinsight](https://github.com/deepinsight): for their [insightface](https://github.com/deepinsight/insightface) project which provided a well-made library and models. Please be reminded that the [use of the model is for non-commercial research purposes only](https://github.com/deepinsight/insightface?tab=readme-ov-file#license).
 -   [havok2-htwo](https://github.com/havok2-htwo): for sharing the code for webcam
 -   [GosuDRM](https://github.com/GosuDRM): for the open version of roop

locales/id.json

@@ -0,0 +1,45 @@
+{
+    "Source x Target Mapper": "Pemetaan Sumber x Target",
+    "select a source image": "Pilih gambar sumber",
+    "Preview": "Pratinjau",
+    "select a target image or video": "Pilih gambar atau video target",
+    "save image output file": "Simpan file keluaran gambar",
+    "save video output file": "Simpan file keluaran video",
+    "select a target image": "Pilih gambar target",
+    "source": "Sumber",
+    "Select a target": "Pilih target",
+    "Select a face": "Pilih wajah",
+    "Keep audio": "Pertahankan audio",
+    "Face Enhancer": "Peningkat wajah",
+    "Many faces": "Banyak wajah",
+    "Show FPS": "Tampilkan FPS",
+    "Keep fps": "Pertahankan FPS",
+    "Keep frames": "Pertahankan frame",
+    "Fix Blueish Cam": "Perbaiki kamera kebiruan",
+    "Mouth Mask": "Masker mulut",
+    "Show Mouth Mask Box": "Tampilkan kotak masker mulut",
+    "Start": "Mulai",
+    "Live": "Langsung",
+    "Destroy": "Hentikan",
+    "Map faces": "Petakan wajah",
+    "Processing...": "Sedang memproses...",
+    "Processing succeed!": "Pemrosesan berhasil!",
+    "Processing ignored!": "Pemrosesan diabaikan!",
+    "Failed to start camera": "Gagal memulai kamera",
+    "Please complete pop-up or close it.": "Harap selesaikan atau tutup pop-up.",
+    "Getting unique faces": "Mengambil wajah unik",
+    "Please select a source image first": "Silakan pilih gambar sumber terlebih dahulu",
+    "No faces found in target": "Tidak ada wajah ditemukan pada target",
+    "Add": "Tambah",
+    "Clear": "Bersihkan",
+    "Submit": "Kirim",
+    "Select source image": "Pilih gambar sumber",
+    "Select target image": "Pilih gambar target",
+    "Please provide mapping!": "Harap tentukan pemetaan!",
+    "At least 1 source with target is required!": "Minimal 1 sumber dengan target diperlukan!",
+    "Face could not be detected in last upload!": "Wajah tidak dapat terdeteksi pada unggahan terakhir!",
+    "Select Camera:": "Pilih Kamera:",
+    "All mappings cleared!": "Semua pemetaan telah dibersihkan!",
+    "Mappings successfully submitted!": "Pemetaan berhasil dikirim!",
+    "Source x Target Mapper is already open.": "Pemetaan Sumber x Target sudah terbuka."
+}

modules/globals.py

@@ -12,7 +12,7 @@ file_types = [
 ]
 
 # Face Mapping Data
-souce_target_map: List[Dict[str, Any]] = [] # Stores detailed map for image/video processing
+source_target_map: List[Dict[str, Any]] = [] # Stores detailed map for image/video processing
 simple_map: Dict[str, Any] = {}             # Stores simplified map (embeddings/faces) for live/simple mode
 
 # Paths
@@ -26,7 +26,7 @@ keep_fps: bool = True
 keep_audio: bool = True
 keep_frames: bool = False
 many_faces: bool = False         # Process all detected faces with default source
-map_faces: bool = False          # Use souce_target_map or simple_map for specific swaps
+map_faces: bool = False          # Use source_target_map or simple_map for specific swaps
 color_correction: bool = False   # Enable color correction (implementation specific)
 nsfw_filter: bool = False
 
@@ -68,4 +68,4 @@ enable_interpolation: bool = True # Toggle temporal smoothing
 interpolation_weight: float = 0  # Blend weight for current frame (0.0-1.0). Lower=smoother.
 # --- END: Added for Frame Interpolation ---
 
-# --- END OF FILE globals.py ---
+# --- END OF FILE globals.py ---

modules/processors/frame/face_swapper.py

@@ -1,8 +1,9 @@
-from typing import Any, List
+from typing import Any, List, Optional
 import cv2
 import insightface
 import threading
 import numpy as np
+import platform
 import modules.globals
 import modules.processors.frame.core
 from modules.core import update_status
@@ -14,9 +15,9 @@ from modules.utilities import (
     is_video,
 )
 from modules.cluster_analysis import find_closest_centroid
-# Removed modules.globals.face_swapper_enabled - assuming controlled elsewhere or implicitly true if used
-# Removed modules.globals.opacity - accessed via getattr
 import os
+from collections import deque
+import time
 
 FACE_SWAPPER = None
 THREAD_LOCK = threading.Lock()
@@ -26,6 +27,16 @@ NAME = "DLC.FACE-SWAPPER"
 PREVIOUS_FRAME_RESULT = None # Stores the final processed frame from the previous step
 # --- END: Added for Interpolation ---
 
+# --- START: Mac M1-M5 Optimizations ---
+IS_APPLE_SILICON = platform.system() == 'Darwin' and platform.machine() == 'arm64'
+FRAME_CACHE = deque(maxlen=3)  # Cache for frame reuse
+FACE_DETECTION_CACHE = {}  # Cache face detections
+LAST_DETECTION_TIME = 0
+DETECTION_INTERVAL = 0.033  # ~30 FPS detection rate for live mode
+FRAME_SKIP_COUNTER = 0
+ADAPTIVE_QUALITY = True
+# --- END: Mac M1-M5 Optimizations ---
+
 abs_dir = os.path.dirname(os.path.abspath(__file__))
 models_dir = os.path.join(
     os.path.dirname(os.path.dirname(os.path.dirname(abs_dir))), "models"
@@ -63,22 +74,40 @@ def get_face_swapper() -> Any:
 
     with THREAD_LOCK:
         if FACE_SWAPPER is None:
-            model_path = os.path.join(models_dir, "inswapper_128_fp16.onnx")
+            model_name = "inswapper_128.onnx"
+            if "CUDAExecutionProvider" in modules.globals.execution_providers:
+                model_name = "inswapper_128_fp16.onnx"
+            model_path = os.path.join(models_dir, model_name)
             update_status(f"Loading face swapper model from: {model_path}", NAME)
             try:
-                # Ensure the providers list is correctly passed
+                # Optimized provider configuration for Apple Silicon
-                providers = modules.globals.execution_providers
+                providers_config = []
-                # print(f"Attempting to load model with providers: {providers}") # Debug print
+                for p in modules.globals.execution_providers:
+                    if p == "CoreMLExecutionProvider" and IS_APPLE_SILICON:
+                        # Enhanced CoreML configuration for M1-M5
+                        providers_config.append((
+                            "CoreMLExecutionProvider",
+                            {
+                                "ModelFormat": "MLProgram",
+                                "MLComputeUnits": "ALL",  # Use Neural Engine + GPU + CPU
+                                "SpecializationStrategy": "FastPrediction",
+                                "AllowLowPrecisionAccumulationOnGPU": 1,
+                                "EnableOnSubgraphs": 1,
+                                "RequireStaticShapes": 0,
+                                "MaximumCacheSize": 1024 * 1024 * 512,  # 512MB cache
+                            }
+                        ))
+                    else:
+                        providers_config.append(p)
+                
                 FACE_SWAPPER = insightface.model_zoo.get_model(
-                    model_path, providers=providers
+                    model_path,
+                    providers=providers_config,
                 )
                 update_status("Face swapper model loaded successfully.", NAME)
             except Exception as e:
                 update_status(f"Error loading face swapper model: {e}", NAME)
-                # print traceback maybe?
+                FACE_SWAPPER = None
-                # import traceback
-                # traceback.print_exc()
-                FACE_SWAPPER = None # Ensure it remains None on failure
                 return None
     return FACE_SWAPPER
 
@@ -87,19 +116,22 @@ def swap_face(source_face: Face, target_face: Face, temp_frame: Frame) -> Frame:
     face_swapper = get_face_swapper()
     if face_swapper is None:
         update_status("Face swapper model not loaded or failed to load. Skipping swap.", NAME)
-        return temp_frame # Return original frame if model failed or not loaded
+        return temp_frame
 
     # Store a copy of the original frame before swapping for opacity blending
     original_frame = temp_frame.copy()
 
-    # --- Pre-swap Input Check (Optional but good practice) ---
+    # Pre-swap Input Check with optimization
     if temp_frame.dtype != np.uint8:
-        # print(f"Warning: Input frame is {temp_frame.dtype}, converting to uint8 before swap.")
         temp_frame = np.clip(temp_frame, 0, 255).astype(np.uint8)
-    # --- End Input Check ---
 
-    # Apply the face swap
+    # Apply the face swap with optimized memory handling
     try:
+        # For Apple Silicon, use optimized inference
+        if IS_APPLE_SILICON:
+            # Ensure contiguous memory layout for better performance
+            temp_frame = np.ascontiguousarray(temp_frame)
+        
         swapped_frame_raw = face_swapper.get(
             temp_frame, target_face, source_face, paste_back=True
         )
@@ -177,14 +209,50 @@ def swap_face(source_face: Face, target_face: Face, temp_frame: Frame) -> Frame:
     return final_swapped_frame
 
 
+# --- START: Mac M1-M5 Optimized Face Detection ---
+def get_faces_optimized(frame: Frame, use_cache: bool = True) -> Optional[List[Face]]:
+    """Optimized face detection for live mode on Apple Silicon"""
+    global LAST_DETECTION_TIME, FACE_DETECTION_CACHE
+    
+    if not use_cache or not IS_APPLE_SILICON:
+        # Standard detection
+        if modules.globals.many_faces:
+            return get_many_faces(frame)
+        else:
+            face = get_one_face(frame)
+            return [face] if face else None
+    
+    # Adaptive detection rate for live mode
+    current_time = time.time()
+    time_since_last = current_time - LAST_DETECTION_TIME
+    
+    # Skip detection if too soon (adaptive frame skipping)
+    if time_since_last < DETECTION_INTERVAL and FACE_DETECTION_CACHE:
+        return FACE_DETECTION_CACHE.get('faces')
+    
+    # Perform detection
+    LAST_DETECTION_TIME = current_time
+    if modules.globals.many_faces:
+        faces = get_many_faces(frame)
+    else:
+        face = get_one_face(frame)
+        faces = [face] if face else None
+    
+    # Cache results
+    FACE_DETECTION_CACHE['faces'] = faces
+    FACE_DETECTION_CACHE['timestamp'] = current_time
+    
+    return faces
+# --- END: Mac M1-M5 Optimized Face Detection ---
+
 # --- START: Helper function for interpolation and sharpening ---
 def apply_post_processing(current_frame: Frame, swapped_face_bboxes: List[np.ndarray]) -> Frame:
-    """Applies sharpening and interpolation."""
+    """Applies sharpening and interpolation with Apple Silicon optimizations."""
     global PREVIOUS_FRAME_RESULT
 
     processed_frame = current_frame.copy()
 
-    # 1. Apply Sharpening (if enabled)
+    # 1. Apply Sharpening (if enabled) with optimized kernel for Apple Silicon
     sharpness_value = getattr(modules.globals, "sharpness", 0.0)
     if sharpness_value > 0.0 and swapped_face_bboxes:
         height, width = processed_frame.shape[:2]
@@ -207,23 +275,21 @@ def apply_post_processing(current_frame: Frame, swapped_face_bboxes: List[np.nda
                 continue
 
             face_region = processed_frame[y1:y2, x1:x2]
-            if face_region.size == 0: continue # Skip empty regions
+            if face_region.size == 0: continue
 
-            # Apply sharpening using addWeighted for smoother control
+            # Apply sharpening with optimized parameters for Apple Silicon
-            # Use try-except for GaussianBlur and addWeighted as they can fail on invalid inputs
             try:
-                 blurred = cv2.GaussianBlur(face_region, (0, 0), 3) # sigma=3, kernel size auto
+                # Use smaller sigma for faster processing on Apple Silicon
-                 sharpened_region = cv2.addWeighted(
+                sigma = 2 if IS_APPLE_SILICON else 3
+                blurred = cv2.GaussianBlur(face_region, (0, 0), sigma)
+                sharpened_region = cv2.addWeighted(
                     face_region, 1.0 + sharpness_value,
                     blurred, -sharpness_value,
                     0
-                 )
+                )
-                 # Ensure the sharpened region doesn't have invalid values
+                sharpened_region = np.clip(sharpened_region, 0, 255).astype(np.uint8)
-                 sharpened_region = np.clip(sharpened_region, 0, 255).astype(np.uint8)
+                processed_frame[y1:y2, x1:x2] = sharpened_region
-                 processed_frame[y1:y2, x1:x2] = sharpened_region
+            except cv2.error:
-            except cv2.error as sharpen_e:
-                # print(f"Warning: OpenCV error during sharpening: {sharpen_e} for bbox {bbox}") # Debug
-                # Skip sharpening for this region if it fails
                 pass
 
 
@@ -323,7 +389,7 @@ def process_frame_v2(temp_frame: Frame, temp_frame_path: str = "") -> Frame:
     source_target_pairs = []
 
     # Ensure maps exist before accessing them
-    souce_target_map = getattr(modules.globals, "souce_target_map", None)
+    source_target_map = getattr(modules.globals, "source_target_map", None)
     simple_map = getattr(modules.globals, "simple_map", None)
 
     # Check if target is a file path (image or video) or live stream
@@ -331,11 +397,11 @@ def process_frame_v2(temp_frame: Frame, temp_frame_path: str = "") -> Frame:
 
     if is_file_target:
         # Processing specific image or video file with pre-analyzed maps
-        if souce_target_map:
+        if source_target_map:
             if modules.globals.many_faces:
                 source_face = default_source_face() # Use default source for all targets
                 if source_face:
-                    for map_data in souce_target_map:
+                    for map_data in source_target_map:
                         if is_image(modules.globals.target_path):
                             target_info = map_data.get("target", {})
                             if target_info: # Check if target info exists
@@ -353,7 +419,7 @@ def process_frame_v2(temp_frame: Frame, temp_frame_path: str = "") -> Frame:
                                          for target_face in faces_in_frame:
                                              source_target_pairs.append((source_face, target_face))
             else: # Single face or specific mapping
-                 for map_data in souce_target_map:
+                 for map_data in source_target_map:
                     source_info = map_data.get("source", {})
                     if not source_info: continue # Skip if no source info
                     source_face = source_info.get("face")

modules/ui.py

@@ -465,7 +465,7 @@ def analyze_target(start: Callable[[], None], root: ctk.CTk):
         return
 
     if modules.globals.map_faces:
-        modules.globals.souce_target_map = []
+        modules.globals.source_target_map = []
 
         if is_image(modules.globals.target_path):
             update_status("Getting unique faces")
@@ -474,8 +474,8 @@ def analyze_target(start: Callable[[], None], root: ctk.CTk):
             update_status("Getting unique faces")
             get_unique_faces_from_target_video()
 
-        if len(modules.globals.souce_target_map) > 0:
+        if len(modules.globals.source_target_map) > 0:
-            create_source_target_popup(start, root, modules.globals.souce_target_map)
+            create_source_target_popup(start, root, modules.globals.source_target_map)
         else:
             update_status("No faces found in target")
     else:
@@ -855,9 +855,9 @@ def webcam_preview(root: ctk.CTk, camera_index: int):
             return
         create_webcam_preview(camera_index)
     else:
-        modules.globals.souce_target_map = []
+        modules.globals.source_target_map = []
         create_source_target_popup_for_webcam(
-            root, modules.globals.souce_target_map, camera_index
+            root, modules.globals.source_target_map, camera_index
         )
 
 

requirements.txt

@@ -11,7 +11,7 @@ tk==0.1.0
 customtkinter==5.2.2
 pillow==11.1.0
 torch; sys_platform != 'darwin'
-torch==2.7.1+cu128; sys_platform == 'darwin'
+torch==2.8.0+cu128; sys_platform == 'darwin'
 torchvision; sys_platform != 'darwin'
 torchvision==0.20.1; sys_platform == 'darwin'
 onnxruntime-silicon==1.16.3; sys_platform == 'darwin' and platform_machine == 'arm64'
@@ -20,4 +20,5 @@ tensorflow; sys_platform != 'darwin'
 opennsfw2==0.10.2
 protobuf==4.25.1
 git+https://github.com/xinntao/BasicSR.git@master
-git+https://github.com/TencentARC/GFPGAN.git@master
+git+https://github.com/TencentARC/GFPGAN.git@master
+pygrabber