Uniform model handling for predictor

facefusion/predictor.py

@@ -1,4 +1,4 @@
-from typing import Any
+from typing import Any, Dict
 import threading
 from functools import lru_cache
 
@@ -9,15 +9,21 @@ from tqdm import tqdm
 
 import facefusion.globals
 from facefusion import wording
-from facefusion.typing import Frame
+from facefusion.typing import Frame, ModelValue
 from facefusion.vision import get_video_frame, count_video_frame_total, read_image, detect_fps
 from facefusion.utilities import resolve_relative_path, conditional_download
 
 PREDICTOR = None
 THREAD_LOCK : threading.Lock = threading.Lock()
 NAME = 'FACEFUSION.PREDICTOR'
-MODEL_URL = 'https://github.com/facefusion/facefusion-assets/releases/download/models/open_nsfw.onnx'
-MODEL_PATH = resolve_relative_path('../.assets/models/open_nsfw.onnx')
+MODELS : Dict[str, ModelValue] =\
+{
+	'open_nsfw':
+	{
+		'url': 'https://github.com/facefusion/facefusion-assets/releases/download/models/open_nsfw.onnx',
+		'path': resolve_relative_path('../.assets/models/open_nsfw.onnx')
+	}
+}
 MAX_PROBABILITY = 0.80
 STREAM_COUNTER = 0
 
@@ -27,7 +33,8 @@ def get_predictor() -> Any:
 
 	with THREAD_LOCK:
 		if PREDICTOR is None:
-			PREDICTOR = onnxruntime.InferenceSession(MODEL_PATH, providers = facefusion.globals.execution_providers)
+			model_path = MODELS.get('open_nsfw').get('path')
+			PREDICTOR = onnxruntime.InferenceSession(model_path, providers = facefusion.globals.execution_providers)
 	return PREDICTOR
 
 
@@ -40,7 +47,8 @@ def clear_predictor() -> None:
 def pre_check() -> bool:
 	if not facefusion.globals.skip_download:
 		download_directory_path = resolve_relative_path('../.assets/models')
-		conditional_download(download_directory_path, [ MODEL_URL ])
+		model_url = MODELS.get('open_nsfw').get('url')
+		conditional_download(download_directory_path, [ model_url ])
 	return True