3.9.0 (#1229)

* mark as next, add AlphaFace model

* add hrffa model

* add hrffa model

* cosmetics in codebase

* more job runner coverage

* bump version 3.9.0

* fix metadata

facefusion/face_landmarker.py

@@ -42,6 +42,32 @@ def create_static_model_set(download_scope : DownloadScope) -> ModelSet:
 			},
 			'size': (256, 256)
 		},
+		'hrffa':
+		{
+			'__metadata__':
+			{
+				'vendor': 'PINTO0309',
+				'license': 'MIT',
+				'year': 2025
+			},
+			'hashes':
+			{
+				'hrffa':
+				{
+					'url': resolve_download_url('models-3.9.0', 'hrffa.hash'),
+					'path': resolve_relative_path('../.assets/models/hrffa.hash')
+				}
+			},
+			'sources':
+			{
+				'hrffa':
+				{
+					'url': resolve_download_url('models-3.9.0', 'hrffa.onnx'),
+					'path': resolve_relative_path('../.assets/models/hrffa.onnx')
+				}
+			},
+			'size': (256, 256)
+		},
 		'peppa_wutz':
 		{
 			'__metadata__':
@@ -119,6 +145,10 @@ def collect_model_downloads() -> Tuple[DownloadSet, DownloadSet]:
 		'fan_68_5': model_set.get('fan_68_5').get('sources').get('fan_68_5')
 	}
 
+	if state_manager.get_item('face_landmarker_model') == 'hrffa':
+		model_hash_set['hrffa'] = model_set.get('hrffa').get('hashes').get('hrffa')
+		model_source_set['hrffa'] = model_set.get('hrffa').get('sources').get('hrffa')
+
 	for face_landmarker_model in [ '2dfan4', 'peppa_wutz' ]:
 		if state_manager.get_item('face_landmarker_model') in [ 'many', face_landmarker_model ]:
 			model_hash_set[face_landmarker_model] = model_set.get(face_landmarker_model).get('hashes').get(face_landmarker_model)
@@ -134,19 +164,21 @@ def pre_check() -> bool:
 
 
 def detect_face_landmark(vision_frame : VisionFrame, bounding_box : BoundingBox, face_angle : Angle) -> Tuple[FaceLandmark68, Score]:
-	face_landmark_2dfan4 = None
-	face_landmark_peppa_wutz = None
-	face_landmark_score_2dfan4 = 0.0
-	face_landmark_score_peppa_wutz = 0.0
+	if state_manager.get_item('face_landmarker_model') == '2dfan4':
+		return detect_with_2dfan4(vision_frame, bounding_box, face_angle)
 
-	if state_manager.get_item('face_landmarker_model') in [ 'many', '2dfan4' ]:
-		face_landmark_2dfan4, face_landmark_score_2dfan4 = detect_with_2dfan4(vision_frame, bounding_box, face_angle)
+	if state_manager.get_item('face_landmarker_model') == 'hrffa':
+		return detect_with_hrffa(vision_frame, bounding_box)
 
-	if state_manager.get_item('face_landmarker_model') in [ 'many', 'peppa_wutz' ]:
-		face_landmark_peppa_wutz, face_landmark_score_peppa_wutz = detect_with_peppa_wutz(vision_frame, bounding_box, face_angle)
+	if state_manager.get_item('face_landmarker_model') == 'peppa_wutz':
+		return detect_with_peppa_wutz(vision_frame, bounding_box, face_angle)
+
+	face_landmark_2dfan4, face_landmark_score_2dfan4 = detect_with_2dfan4(vision_frame, bounding_box, face_angle)
+	face_landmark_peppa_wutz, face_landmark_score_peppa_wutz = detect_with_peppa_wutz(vision_frame, bounding_box, face_angle)
 
 	if face_landmark_score_2dfan4 > face_landmark_score_peppa_wutz - 0.2:
 		return face_landmark_2dfan4, face_landmark_score_2dfan4
+
 	return face_landmark_peppa_wutz, face_landmark_score_peppa_wutz
 
 
@@ -155,10 +187,12 @@ def detect_with_2dfan4(temp_vision_frame: VisionFrame, bounding_box: BoundingBox
 	scale = 195 / numpy.subtract(bounding_box[2:], bounding_box[:2]).max().clip(1, None)
 	translation = (model_size[0] - numpy.add(bounding_box[2:], bounding_box[:2]) * scale) * 0.5
 	rotation_matrix, rotation_size = create_rotation_matrix_and_size(face_angle, model_size)
+
 	crop_vision_frame, affine_matrix = warp_face_by_translation(temp_vision_frame, translation, scale, model_size)
 	crop_vision_frame = cv2.warpAffine(crop_vision_frame, rotation_matrix, rotation_size)
 	crop_vision_frame = conditional_optimize_contrast(crop_vision_frame)
 	crop_vision_frame = crop_vision_frame.transpose(2, 0, 1).astype(numpy.float32) / 255.0
+
 	face_landmark_68, face_heatmap = forward_with_2dfan4(crop_vision_frame)
 	face_landmark_68 = face_landmark_68[:, :, :2][0] / 64 * 256
 	face_landmark_68 = transform_points(face_landmark_68, cv2.invertAffineTransform(rotation_matrix))
@@ -166,6 +200,25 @@ def detect_with_2dfan4(temp_vision_frame: VisionFrame, bounding_box: BoundingBox
 	face_landmark_score_68 = numpy.amax(face_heatmap, axis = (2, 3))
 	face_landmark_score_68 = numpy.mean(face_landmark_score_68)
 	face_landmark_score_68 = numpy.interp(face_landmark_score_68, [ 0, 0.9 ], [ 0, 1 ])
+
+	return face_landmark_68, face_landmark_score_68
+
+
+def detect_with_hrffa(temp_vision_frame : VisionFrame, bounding_box : BoundingBox) -> Tuple[FaceLandmark68, Score]:
+	model_size = create_static_model_set('full').get('hrffa').get('size')
+	scale = model_size[0] / (numpy.subtract(bounding_box[2:], bounding_box[:2]).max().clip(1, None) * 1.7)
+	translation = (model_size[0] - numpy.add(bounding_box[2:], bounding_box[:2]) * scale) * 0.5
+
+	crop_vision_frame, affine_matrix = warp_face_by_translation(temp_vision_frame, translation, scale, model_size)
+	crop_vision_frame = crop_vision_frame[:, :, ::-1].transpose(2, 0, 1).astype(numpy.float32) / 255.0
+	crop_vision_frame = (crop_vision_frame - 0.5) / 0.5
+	crop_vision_frame = numpy.expand_dims(crop_vision_frame, axis = 0)
+
+	face_landmark_68 = forward_with_hrffa(crop_vision_frame)
+	face_landmark_68 = face_landmark_68.reshape(-1, 2) * model_size[0]
+	face_landmark_68 = transform_points(face_landmark_68, cv2.invertAffineTransform(affine_matrix))
+	face_landmark_score_68 = 1.0
+
 	return face_landmark_68, face_landmark_score_68
 
 
@@ -174,17 +227,20 @@ def detect_with_peppa_wutz(temp_vision_frame : VisionFrame, bounding_box : Bound
 	scale = 195 / numpy.subtract(bounding_box[2:], bounding_box[:2]).max().clip(1, None)
 	translation = (model_size[0] - numpy.add(bounding_box[2:], bounding_box[:2]) * scale) * 0.5
 	rotation_matrix, rotation_size = create_rotation_matrix_and_size(face_angle, model_size)
+
 	crop_vision_frame, affine_matrix = warp_face_by_translation(temp_vision_frame, translation, scale, model_size)
 	crop_vision_frame = cv2.warpAffine(crop_vision_frame, rotation_matrix, rotation_size)
 	crop_vision_frame = conditional_optimize_contrast(crop_vision_frame)
 	crop_vision_frame = crop_vision_frame.transpose(2, 0, 1).astype(numpy.float32) / 255.0
 	crop_vision_frame = numpy.expand_dims(crop_vision_frame, axis = 0)
+
 	prediction = forward_with_peppa_wutz(crop_vision_frame)
 	face_landmark_68 = prediction.reshape(-1, 3)[:, :2] / 64 * model_size[0]
 	face_landmark_68 = transform_points(face_landmark_68, cv2.invertAffineTransform(rotation_matrix))
 	face_landmark_68 = transform_points(face_landmark_68, cv2.invertAffineTransform(affine_matrix))
 	face_landmark_score_68 = prediction.reshape(-1, 3)[:, 2].mean()
 	face_landmark_score_68 = numpy.interp(face_landmark_score_68, [ 0, 0.95 ], [ 0, 1 ])
+
 	return face_landmark_68, face_landmark_score_68
 
 
@@ -216,6 +272,18 @@ def forward_with_2dfan4(crop_vision_frame : VisionFrame) -> Tuple[Prediction, Pr
 	return prediction
 
 
+def forward_with_hrffa(crop_vision_frame : VisionFrame) -> Prediction:
+	face_landmarker = get_inference_pool().get('hrffa')
+
+	with conditional_thread_semaphore():
+		prediction = face_landmarker.run(None,
+		{
+			'input': crop_vision_frame
+		})[0]
+
+	return prediction
+
+
 def forward_with_peppa_wutz(crop_vision_frame : VisionFrame) -> Prediction:
 	face_landmarker = get_inference_pool().get('peppa_wutz')
 

facefusion/jobs/job_manager.py

@@ -10,7 +10,7 @@ from facefusion.sanitizer import sanitize_job_id
 from facefusion.time_helper import get_current_date_time
 from facefusion.types import Args, Job, JobSet, JobStatus, JobStep, JobStepStatus
 
-JOBS_PATH : Optional[str] = None
+JOBS_PATH : str = '.jobs'
 
 
 def init_jobs(jobs_path : str) -> bool:

facefusion/metadata.py

@@ -4,7 +4,7 @@ METADATA =\
 {
 	'name': 'FaceFusion',
 	'description': 'Industry leading face manipulation platform',
-	'version': '3.8.3',
+	'version': '3.9.0',
 	'license': 'OpenRAIL-AS',
 	'author': 'Henry Ruhs',
 	'url': 'https://facefusion.io'

facefusion/processors/modules/background_remover/core.py

@@ -122,7 +122,7 @@ def create_static_model_set(download_scope : DownloadScope) -> ModelSet:
 			{
 				'vendor': 'nikopueringer',
 				'license': 'Non-Commercial',
-				'year': 2025
+				'year': 2026
 			},
 			'hashes':
 			{
@@ -151,7 +151,7 @@ def create_static_model_set(download_scope : DownloadScope) -> ModelSet:
 			{
 				'vendor': 'nikopueringer',
 				'license': 'Non-Commercial',
-				'year': 2025
+				'year': 2026
 			},
 			'hashes':
 			{

facefusion/processors/modules/face_swapper/choices.py

@@ -6,6 +6,7 @@ from facefusion.processors.modules.face_swapper.types import FaceSwapperModel, F
 
 face_swapper_set : FaceSwapperSet =\
 {
+	'alphaface_256': [ '256x256', '512x512', '768x768', '1024x1024' ],
 	'blendswap_256': [ '256x256', '384x384', '512x512', '768x768', '1024x1024' ],
 	'ghost_1_256': [ '256x256', '512x512', '768x768', '1024x1024' ],
 	'ghost_2_256': [ '256x256', '512x512', '768x768', '1024x1024' ],

facefusion/processors/modules/face_swapper/core.py

@@ -33,6 +33,36 @@ from facefusion.vision import read_static_image, read_static_images, read_static
 def create_static_model_set(download_scope : DownloadScope) -> ModelSet:
 	return\
 	{
+		'alphaface_256':
+		{
+			'__metadata__':
+			{
+				'vendor': 'AlphaFace',
+				'license': 'Non-Commercial',
+				'year': 2026
+			},
+			'hashes':
+			{
+				'face_swapper':
+				{
+					'url': resolve_download_url('models-3.9.0', 'alphaface_256.hash'),
+					'path': resolve_relative_path('../.assets/models/alphaface_256.hash')
+				}
+			},
+			'sources':
+			{
+				'face_swapper':
+				{
+					'url': resolve_download_url('models-3.9.0', 'alphaface_256.onnx'),
+					'path': resolve_relative_path('../.assets/models/alphaface_256.onnx')
+				}
+			},
+			'type': 'alphaface',
+			'template': 'arcface_128',
+			'size': (256, 256),
+			'mean': [ 0.0, 0.0, 0.0 ],
+			'standard_deviation': [ 1.0, 1.0, 1.0 ]
+		},
 		'blendswap_256':
 		{
 			'__metadata__':
@@ -689,6 +719,10 @@ def prepare_source_frame(source_face : Face, source_vision_frame : VisionFrame)
 def prepare_source_embedding(source_face : Face) -> Embedding:
 	model_type = get_model_options().get('type')
 
+	if model_type == 'alphaface':
+		source_embedding = source_face.embedding.reshape((1, -1))
+		return source_embedding
+
 	if model_type == 'ghost':
 		source_embedding = source_face.embedding.reshape(-1, 512)
 		source_embedding, _ = convert_source_embedding(source_embedding)
@@ -717,7 +751,7 @@ def balance_source_embedding(source_embedding : Embedding, target_embedding : Em
 	face_swapper_weight = state_manager.get_item('face_swapper_weight')
 	face_swapper_weight = numpy.interp(face_swapper_weight, [ 0, 1 ], [ 0.35, -0.35 ]).astype(numpy.float32)
 
-	if model_type in [ 'hififace', 'hyperswap', 'inswapper', 'simswap' ]:
+	if model_type in [ 'alphaface', 'hififace', 'hyperswap', 'inswapper', 'simswap' ]:
 		target_embedding = target_embedding / numpy.linalg.norm(target_embedding)
 
 	source_embedding = source_embedding.reshape(1, -1)

facefusion/processors/modules/face_swapper/types.py

@@ -11,7 +11,7 @@ FaceSwapperInputs = TypedDict('FaceSwapperInputs',
 	'temp_vision_mask' : Mask
 })
 
-FaceSwapperModel = Literal['blendswap_256', 'ghost_1_256', 'ghost_2_256', 'ghost_3_256', 'hififace_unofficial_256', 'hyperswap_1a_256', 'hyperswap_1b_256', 'hyperswap_1c_256', 'inswapper_128', 'inswapper_128_fp16', 'simswap_256', 'simswap_unofficial_512', 'uniface_256']
+FaceSwapperModel = Literal['alphaface_256', 'blendswap_256', 'ghost_1_256', 'ghost_2_256', 'ghost_3_256', 'hififace_unofficial_256', 'hyperswap_1a_256', 'hyperswap_1b_256', 'hyperswap_1c_256', 'inswapper_128', 'inswapper_128_fp16', 'simswap_256', 'simswap_unofficial_512', 'uniface_256']
 
 FaceSwapperWeight : TypeAlias = float
 

facefusion/types.py

@@ -175,7 +175,7 @@ TableHeader : TypeAlias = str
 TableContent : TypeAlias = Any
 
 FaceDetectorModel = Literal['many', 'retinaface', 'scrfd', 'yolo_face', 'yunet']
-FaceLandmarkerModel = Literal['many', '2dfan4', 'peppa_wutz']
+FaceLandmarkerModel = Literal['many', '2dfan4', 'hrffa', 'peppa_wutz']
 FaceDetectorSet : TypeAlias = Dict[FaceDetectorModel, List[str]]
 FaceSelectorMode = Literal['many', 'one', 'reference']
 FaceSelectorOrder = Literal['left-right', 'right-left', 'top-bottom', 'bottom-top', 'small-large', 'large-small', 'best-worst', 'worst-best']

tests/test_job_runner.py

@@ -5,7 +5,7 @@ from facefusion import ffmpeg, ffmpeg_builder, process_manager
 from facefusion.download import conditional_download
 from facefusion.filesystem import copy_file
 from facefusion.jobs.job_manager import add_step, clear_jobs, create_job, init_jobs, move_job_file, submit_job, submit_jobs
-from facefusion.jobs.job_runner import collect_output_set, finalize_steps, retry_job, retry_jobs, run_job, run_jobs, run_steps
+from facefusion.jobs.job_runner import clean_steps, collect_output_set, finalize_steps, retry_job, retry_jobs, run_job, run_jobs, run_steps
 from facefusion.types import Args
 from .helper import get_test_example_file, get_test_examples_directory, get_test_jobs_directory, get_test_output_file, is_test_output_file, prepare_test_output_directory
 
@@ -241,6 +241,35 @@ def test_finalize_steps() -> None:
 	assert is_test_output_file('output-3.jpg') is True
 
 
+def test_clean_steps() -> None:
+	args_1 =\
+	{
+		'source_path': get_test_example_file('source.jpg'),
+		'target_path': get_test_example_file('target-240p.mp4'),
+		'output_path': get_test_output_file('output-1.mp4')
+	}
+	args_2 =\
+	{
+		'source_path': get_test_example_file('source.jpg'),
+		'target_path': get_test_example_file('target-240p.jpg'),
+		'output_path': get_test_output_file('output-2.jpg')
+	}
+
+	create_job('job-test-clean-steps')
+	add_step('job-test-clean-steps', args_1)
+	add_step('job-test-clean-steps', args_2)
+
+	copy_file(args_1.get('target_path'), get_test_output_file('output-1-job-test-clean-steps-0.mp4'))
+	copy_file(args_2.get('target_path'), get_test_output_file('output-2-job-test-clean-steps-1.jpg'))
+
+	assert is_test_output_file('output-1-job-test-clean-steps-0.mp4') is True
+	assert is_test_output_file('output-2-job-test-clean-steps-1.jpg') is True
+
+	assert clean_steps('job-test-clean-steps') is True
+	assert is_test_output_file('output-1-job-test-clean-steps-0.mp4') is False
+	assert is_test_output_file('output-2-job-test-clean-steps-1.jpg') is False
+
+
 def test_collect_output_set() -> None:
 	args_1 =\
 	{