{"id":13718,"date":"2024-02-29T16:37:32","date_gmt":"2024-02-29T08:37:32","guid":{"rendered":"https:\/\/92it.top\/?p=13718"},"modified":"2024-04-06T17:11:59","modified_gmt":"2024-04-06T09:11:59","slug":"%e8%af%a6%e8%a7%a3-facefusion-2-3-0-%e7%9a%84%e5%8f%82%e6%95%b0","status":"publish","type":"post","link":"https:\/\/92it.top\/?p=13718","title":{"rendered":"\u8be6\u89e3 FaceFusion (2.3.0) \u7684\u53c2\u6570"},"content":{"rendered":"\n

\u539f\u6587\u94fe\u63a5\uff1ahttps:\/\/blog.csdn.net\/ddrfan\/article\/details\/136226323<\/a><\/p>\n\n\n\n

\u7279\u70b9\u548c\u754c\u9762<\/strong><\/p>\n\n\n\n

\n\n\n\n

\u542f\u52a8WEBUI\u540e\uff0c\u5927\u6982\u662f\u8fd9\u4e2a\u6837\u5b50\u3002
\u770b\u8d77\u6765\u6709\u70b9\u590d\u6742\uff0c\u6ca1\u6709\u6700\u65b0\u7684Rope\u90a3\u4e48\u7b80\u6d01\u3002
\u548cRope\u7684\u5bf9\u6bd4\uff0c\u53ef\u4ee5\u770b\u4e0a\u6b21\u5199\u7684<\/a> https:\/\/92it.top\/?p=13706<\/a><\/p>\n\n\n\n

\u4f46\u5176\u5b9e\u628a\u4e00\u4e9b\u53c2\u6570\u8bbe\u7f6e<\/a>\u597d\u540e\uff0c\u6bcf\u6b21\u4f7f\u7528\u4e5f\u662f\u76f8\u5f53\u5feb\u6377\u7684\u3002<\/p>\n\n\n

\n
\"\"<\/figure><\/div>\n\n\n

\uff08\u4e00\uff09\u53c2\u6570\u63a7\u5236\u7684FaceFusion<\/strong><\/p>\n\n\n\n

\n\n\n\n

\u524d\u9762\u63d0\u5230\u8fc7\uff0cFaceFusion\u548c\u4fa7\u91cdGUI\u4ea4\u4e92\u7684Rope\u4e0d\u4e00\u6837\uff0c\u53ef\u4ee5\u65e0\u754c\u9762\u8fdb\u884c\u547d\u4ee4\u884c\u64cd\u4f5c(CLI)\u3002
\u6240\u4ee5\u5b83\u6709\u975e\u5e38\u975e\u5e38\u591a\u7684\u53c2\u6570\u53ef\u9009\u3002
\u5f53\u7136\uff0c\u5982\u679c\u4f60\u542f\u52a8\u7684\u662fWEBUI\uff0c\u90a3\u4e48\u5927\u90e8\u5206\u53c2\u6570\u4e5f\u53ef\u4ee5\u901a\u8fc7\u754c\u9762\u8c03\u6574\u3002
\u800c\u5176\u4e2d\u5c0f\u90e8\u5206\u53c2\u6570\u5fc5\u987b\u5728\u547d\u4ee4\u884c\u4e2d\u751f\u6548\u3002<\/p>\n\n\n\n

\uff081.1\uff09\u542f\u52a8\u53c2\u6570<\/strong><\/p>\n\n\n\n

\u5efa\u8bae\u914d\u5408\u5b98\u65b9\u5e2e\u52a9<\/a>\u6765\u4f7f\u7528\u548c\u4e86\u89e3\u53c2\u6570\u542b\u4e49\u3002
\u90a3\u4e48\u76ee\u524d\u7248\u672c\u6709\u8fd9\u4e9b\u53c2\u6570\u3002<\/p>\n\n\n\n

python run.py [options]\n\noptions:\n  -h, --help                                                                                                             show this help message and exit\n  -s SOURCE_PATHS, --source SOURCE_PATHS                                                                                 choose single or multiple source images or audios\n  -t TARGET_PATH, --target TARGET_PATH                                                                                   choose single target image or video\n  -o OUTPUT_PATH, --output OUTPUT_PATH                                                                                   specify the output file or directory\n  -v, --version                                                                                                          show program's version number and exit\n\nmisc:\n  --skip-download                                                                                                        omit automate downloads and remote lookups\n  --headless                                                                                                             run the program without a user interface\n  --log-level {error,warn,info,debug}                                                                                    adjust the message severity displayed in the terminal\n\nexecution:\n  --execution-providers EXECUTION_PROVIDERS [EXECUTION_PROVIDERS ...]                                                    accelerate the model inference using different providers (choices: cpu, ...)\n  --execution-thread-count [1-128]                                                                                       specify the amount of parallel threads while processing\n  --execution-queue-count [1-32]                                                                                         specify the amount of frames each thread is processing\n\nmemory:\n  --video-memory-strategy {strict,moderate,tolerant}                                                                     balance fast frame processing and low vram usage\n  --system-memory-limit [0-128]                                                                                          limit the available ram that can be used while processing\n\nface analyser:\n  --face-analyser-order {left-right,right-left,top-bottom,bottom-top,small-large,large-small,best-worst,worst-best}      specify the order in which the face analyser detects faces.\n  --face-analyser-age {child,teen,adult,senior}                                                                          filter the detected faces based on their age\n  --face-analyser-gender {female,male}                                                                                   filter the detected faces based on their gender\n  --face-detector-model {retinaface,yoloface,yunet}                                                                      choose the model responsible for detecting the face\n  --face-detector-size FACE_DETECTOR_SIZE                                                                                specify the size of the frame provided to the face detector\n  --face-detector-score [0.0-1.0]                                                                                        filter the detected faces base on the confidence score\n\nface selector:\n  --face-selector-mode {reference,one,many}                                                                              use reference based tracking with simple matching\n  --reference-face-position REFERENCE_FACE_POSITION                                                                      specify the position used to create the reference face\n  --reference-face-distance [0.0-1.5]                                                                                    specify the desired similarity between the reference face and target face\n  --reference-frame-number REFERENCE_FRAME_NUMBER                                                                        specify the frame used to create the reference face\n\nface mask:\n  --face-mask-types FACE_MASK_TYPES [FACE_MASK_TYPES ...]                                                                mix and match different face mask types (choices: box, occlusion, region)\n  --face-mask-blur [0.0-1.0]                                                                                             specify the degree of blur applied the box mask\n  --face-mask-padding FACE_MASK_PADDING [FACE_MASK_PADDING ...]                                                          apply top, right, bottom and left padding to the box mask\n  --face-mask-regions FACE_MASK_REGIONS [FACE_MASK_REGIONS ...]                                                          choose the facial features used for the region mask (choices: skin, left-eyebrow, right-eyebrow, left-eye, right-eye, eye-glasses, nose, mouth, upper-lip, lower-lip)\n\nframe extraction:\n  --trim-frame-start TRIM_FRAME_START                                                                                    specify the the start frame of the target video\n  --trim-frame-end TRIM_FRAME_END                                                                                        specify the the end frame of the target video\n  --temp-frame-format {bmp,jpg,png}                                                                                      specify the temporary resources format\n  --temp-frame-quality [0-100]                                                                                           specify the temporary resources quality\n  --keep-temp                                                                                                            keep the temporary resources after processing\n\noutput creation:\n  --output-image-quality [0-100]                                                                                         specify the image quality which translates to the compression factor\n  --output-video-encoder {libx264,libx265,libvpx-vp9,h264_nvenc,hevc_nvenc}                                              specify the encoder use for the video compression\n  --output-video-preset {ultrafast,superfast,veryfast,faster,fast,medium,slow,slower,veryslow}                           balance fast video processing and video file size\n  --output-video-quality [0-100]                                                                                         specify the video quality which translates to the compression factor\n  --output-video-resolution OUTPUT_VIDEO_RESOLUTION                                                                      specify the video output resolution based on the target video\n  --output-video-fps OUTPUT_VIDEO_FPS                                                                                    specify the video output fps based on the target video\n  --skip-audio                                                                                                           omit the audio from the target video\n\nframe processors:\n  --frame-processors FRAME_PROCESSORS [FRAME_PROCESSORS ...]                                                             load a single or multiple frame processors. (choices: face_debugger, face_enhancer, face_swapper, frame_enhancer, lip_syncer, ...)\n  --face-debugger-items FACE_DEBUGGER_ITEMS [FACE_DEBUGGER_ITEMS ...]                                                    load a single or multiple frame processors (choices: bounding-box, landmark-5, landmark-68, face-mask, score, age, gender)\n  --face-enhancer-model {codeformer,gfpgan_1.2,gfpgan_1.3,gfpgan_1.4,gpen_bfr_256,gpen_bfr_512,restoreformer_plus_plus}  choose the model responsible for enhancing the face\n  --face-enhancer-blend [0-100]                                                                                          blend the enhanced into the previous face\n  --face-swapper-model {blendswap_256,inswapper_128,inswapper_128_fp16,simswap_256,simswap_512_unofficial,uniface_256}   choose the model responsible for swapping the face\n  --frame-enhancer-model {real_esrgan_x2plus,real_esrgan_x4plus,real_esrnet_x4plus}                                      choose the model responsible for enhancing the frame\n  --frame-enhancer-blend [0-100]                                                                                         blend the enhanced into the previous frame\n  --lip-syncer-model {wav2lip_gan}                                                                                       choose the model responsible for syncing the lips\n\nuis:\n  --ui-layouts UI_LAYOUTS [UI_LAYOUTS ...]                                                                               launch a single or multiple UI layouts (choices: benchmark, default, webcam, ...)\n<\/pre>\n\n\n\n
\u6bd4\u5982\u4e0b\u9762\u8fd9\u4e2a\u4f8b\u5b50\u3002
\u6307\u5b9a\u4e86\u8f93\u51fa\u76ee\u5f55\uff0c\u90e8\u5206\u6a21\u578b\uff0c\u89c6\u9891\u7f16\u7801\u7b49\uff1a<\/p>\n\n\n\n
python run.py -o D:\\Temp\\out\\A --skip-download --execution-providers cuda --frame-processors face_swapper face_enhancer --face-enhancer-model gpen_bfr_256 --face-enhancer-blend 90 --output-video-encoder hevc_nvenc --output-video-quality 99 --face-mask-types box occlusion\r<\/pre>\n\n\n\n
\uff081.2\uff09\u914d\u7f6e\u6587\u4ef6\u8986\u76d6<\/strong><\/p>\n\n\n\n
\u4e0d\u60f3\u6bcf\u6b21\u90fd\u6253\u5f88\u957f\u7684\u547d\u4ee4\u884c\uff0c\u53ef\u4ee5\u5199\u5165\u4e00\u4e2abat\u6587\u4ef6\u4e2d\u3002<\/p>\n\n\n\n
\u7136\u540e\u4f5c\u8005\u8003\u8651\u5230\u4e86\u8fd9\u70b9\uff0c\u6709\u4e2a\u66f4\u7b80\u5355\u7684\u529e\u6cd5\u3002
\u628a\u8fd9\u4e9b\u53c2\u6570\u5199\u5165facefusion.ini\u914d\u7f6e\u6587\u4ef6\uff0c\u53ef\u4ee5\u8d77\u5230\u547d\u4ee4\u884c\u53c2\u6570\u540c\u6837\u7684\u4f5c\u7528\uff0c\u540c\u65f6\u66f4\u6574\u6d01\u6613\u8bfb\u3002<\/p>\n\n\n\n
\u4f46\u9879\u76ee\u4e2d\u4f5c\u8005\u63d0\u4f9b\u7684\u662ffacefusion.ini\u672c\u8eab\uff0c\u800c\u4e0d\u662ffacefusion.ini.sample\uff0c\u4f1a\u5f71\u54cd\u62c9\u4ee3\u7801\u554a\u3002<\/p>\n\n\n\n
[general]\nsource_paths =\ntarget_path =\noutput_path =\n\n[misc]\nskip_download =\nheadless =\nlog_level =\n\n[execution]\nexecution_providers =\nexecution_thread_count =\nexecution_queue_count =\n\n[memory]\nvideo_memory_strategy =\nsystem_memory_limit =\n\n[face_analyser]\nface_analyser_order =\nface_analyser_age =\nface_analyser_gender =\nface_detector_model =\nface_detector_size =\nface_detector_score =\n\n[face_selector]\nface_selector_mode =\nreference_face_position =\nreference_face_distance =\nreference_frame_number =\n\n[face_mask]\nface_mask_types =\nface_mask_blur =\nface_mask_padding =\nface_mask_regions =\n\n[frame_extraction]\ntrim_frame_start =\ntrim_frame_end =\ntemp_frame_format =\ntemp_frame_quality =\nkeep_temp =\n\n[output_creation]\noutput_image_quality =\noutput_video_encoder =\noutput_video_preset =\noutput_video_quality =\noutput_video_resolution =\noutput_video_fps =\nskip_audio =\n\n[frame_processors]\nframe_processors =\nface_debugger_items =\nface_enhancer_model =\nface_enhancer_blend =\nface_swapper_model =\nframe_enhancer_model =\nframe_enhancer_blend =\nlip_syncer_model =\n\n[uis]\nui_layouts =\n<\/pre>\n\n\n\n
\uff08\u4e8c\uff09\u53c2\u6570\u8be6\u89e3<\/strong><\/p>\n\n\n\n
\n\n\n\n
\uff082.1\uff09UI\u53c2\u6570<\/strong><\/p>\n\n\n\n
\uff082.1.1\uff09UI\u5e03\u5c40<\/strong><\/p>\n\n\n\n
\ud83d\udca1\u4fdd\u6301\u9ed8\u8ba4\u5c31\u597d\uff0c\u54b1\u901a\u5e38\u7528\u4e0d\u5230\u3002<\/p>\n\n\n\n
\u7528\u6765\u5207\u6362\u4e0d\u540c\u7684UI\u5e03\u5c40\u3002
\u7528\u6cd5: –ui-layouts<\/p>\n\n\n\n
\u9ed8\u8ba4\u503c: [ default ]\uff0c\u53ef\u9009\u8fd9\u4e9b\u503c\uff1a<\/p>\n\n\n\n
benchmark \uff08\u57fa\u51c6\u6d4b\u8bd5\uff09\ndefault\uff08\u9ed8\u8ba4\u503c\uff0c\u6362\u89c6\u9891\uff09\nwebcam \uff08\u6444\u50cf\u5934\uff0c\u76f4\u64ad\uff09<\/pre>\n\n\n\n
\u6bd4\u5982: –ui-layouts benchmark \u5c31\u542f\u52a8\u6d4b\u8bd5\u6027\u80fd\u7684UI\u5e03\u5c40\u3002<\/p>\n\n\n\n
\uff082.2\uff09\u6742\u9879\u53c2\u6570<\/strong><\/p>\n\n\n\n
\uff082.2.1\uff09Skip Download<\/strong><\/p>\n\n\n\n
\ud83d\udca1\u5929\u671d\u7f51\u7edc\u5fc5\u987b\u5f97\u5f00\uff08\u9664\u975e\u4f60\u7684\u7f51\u7edc\u81ea\u52a8\u4e0b\u8f7d\u6a21\u578b\u6ca1\u95ee\u9898\uff09\u3002<\/p>\n\n\n\n
\u987e\u540d\u601d\u4e49\u7528\u6765\u8df3\u8fc7\u8fdc\u7a0b\u4e0b\u8f7d\uff08\u5fc5\u8981\u7684\u6a21\u578b\u6211\u4eec\u53ef\u4ee5\u901a\u8fc7\u522b\u7684\u65b9\u5f0f\uff0c\u63d0\u524d\u4e0b\u8f7d\u597d\uff09
\u7528\u6cd5: --skip-download<\/code><\/p>\n\n\n\n
\uff082.2.2\uff09Headless<\/strong><\/p>\n\n\n\n
\ud83d\udca1\u4fdd\u6301\u9ed8\u8ba4\u5c31\u597d\uff0c\u9664\u975e\u2026\u2026<\/p>\n\n\n\n
\u4e0d\u4f7f\u7528\u754c\u9762\u7684\u65b9\u5f0f\u542f\u52a8\u7a0b\u5e8f\uff0c\u9002\u5408\u5728\u670d\u52a1\u5668\u4e0a\u64cd\u4f5c\u3002
\u8bbe\u7f6e\u597d\u5404\u79cd\u53c2\u6570\u540e\uff0c\u4e00\u6761\u547d\u4ee4\u8fd0\u884c\u76f4\u63a5\u51fa\u7ed3\u679c\u3002
\u7528\u6cd5: --headless<\/code><\/p>\n\n\n\n
\uff082.2.3\uff09Log Level<\/strong><\/p>\n\n\n\n
\ud83d\udca1\u4fdd\u6301\u9ed8\u8ba4\u5c31\u597d\uff0c\u54b1\u901a\u5e38\u7528\u4e0d\u5230\u3002<\/p>\n\n\n\n
\u8c03\u6574\u63a7\u5236\u53f0\u8f93\u51fa\u7684\u65e5\u5fd7\u7ea7\u522b\u3002
\u7528\u6cd5: --log-level<\/code><\/p>\n\n\n\n
\u9ed8\u8ba4\u503c: info<\/code>\uff0c\u53ef\u9009\u8fd9\u4e9b\u503c\uff1a<\/p>\n\n\n\n