diff --git a/vibe_voice_tts/vibe_voice_tts.json b/vibe_voice_tts/vibe_voice_tts.json new file mode 100644 index 0000000..8ff9f68 --- /dev/null +++ b/vibe_voice_tts/vibe_voice_tts.json @@ -0,0 +1,376 @@ +{ + "id": "dea7c2e7-debc-4d86-9663-ffc7bce608f4", + "revision": 0, + "last_node_id": 14, + "last_link_id": 10, + "nodes": [ + { + "id": 5, + "type": "SaveAudioMP3", + "pos": [ + 1155.41943359375, + 669.2053833007812 + ], + "size": [ + 270, + 136 + ], + "flags": {}, + "order": 7, + "mode": 0, + "inputs": [ + { + "name": "audio", + "type": "AUDIO", + "link": 4 + } + ], + "outputs": [], + "properties": { + "cnr_id": "comfy-core", + "ver": "0.3.50", + "Node name for S&R": "SaveAudioMP3" + }, + "widgets_values": [ + "audio/ComfyUI", + "320k" + ] + }, + { + "id": 8, + "type": "LoadAudio", + "pos": [ + 374.70074462890625, + 833.9228515625 + ], + "size": [ + 274.4583435058594, + 136 + ], + "flags": {}, + "order": 0, + "mode": 4, + "inputs": [], + "outputs": [ + { + "name": "AUDIO", + "type": "AUDIO", + "links": [ + 8 + ] + } + ], + "properties": { + "cnr_id": "comfy-core", + "ver": "0.3.50", + "Node name for S&R": "LoadAudio" + }, + "widgets_values": [ + "ComfyUI_00009_.mp4", + null, + null + ] + }, + { + "id": 3, + "type": "PreviewAudio", + "pos": [ + 1159.3675537109375, + 370.2304992675781 + ], + "size": [ + 270, + 88 + ], + "flags": {}, + "order": 6, + "mode": 0, + "inputs": [ + { + "name": "audio", + "type": "AUDIO", + "link": 2 + } + ], + "outputs": [], + "properties": { + "cnr_id": "comfy-core", + "ver": "0.3.50", + "Node name for S&R": "PreviewAudio" + }, + "widgets_values": [] + }, + { + "id": 2, + "type": "LoadAudio", + "pos": [ + 376.3929443359375, + 248.9485321044922 + ], + "size": [ + 274.4583435058594, + 136 + ], + "flags": {}, + "order": 1, + "mode": 0, + "inputs": [], + "outputs": [ + { + "name": "AUDIO", + "type": "AUDIO", + "links": [ + 1 + ] + } + ], + "properties": { + "cnr_id": "comfy-core", + "ver": "0.3.50", + "Node name for S&R": "LoadAudio" + }, + "widgets_values": [ + "ai_rncz_real_voice.wav", + null, + null + ] + }, + { + "id": 7, + "type": "LoadAudio", + "pos": [ + 375.828857421875, + 643.8200073242188 + ], + "size": [ + 274.4583435058594, + 136 + ], + "flags": { + "collapsed": false + }, + "order": 2, + "mode": 4, + "inputs": [], + "outputs": [ + { + "name": "AUDIO", + "type": "AUDIO", + "links": [ + 6 + ] + } + ], + "properties": { + "cnr_id": "comfy-core", + "ver": "0.3.50", + "Node name for S&R": "LoadAudio" + }, + "widgets_values": [ + "ComfyUI_00009_.mp4", + null, + null + ] + }, + { + "id": 10, + "type": "LoadAudio", + "pos": [ + 371.3160400390625, + 441.87164306640625 + ], + "size": [ + 274.4583435058594, + 136 + ], + "flags": { + "collapsed": false + }, + "order": 3, + "mode": 4, + "inputs": [], + "outputs": [ + { + "name": "AUDIO", + "type": "AUDIO", + "links": [ + 10 + ] + } + ], + "properties": { + "cnr_id": "comfy-core", + "ver": "0.3.50", + "Node name for S&R": "LoadAudio" + }, + "widgets_values": [ + "ComfyUI_00009_.mp4", + null, + null + ] + }, + { + "id": 1, + "type": "VibeVoiceTTS", + "pos": [ + 706.9571533203125, + 367.4100341796875 + ], + "size": [ + 400, + 436 + ], + "flags": {}, + "order": 5, + "mode": 0, + "inputs": [ + { + "name": "speaker_1_voice", + "shape": 7, + "type": "AUDIO", + "link": 1 + }, + { + "name": "speaker_2_voice", + "shape": 7, + "type": "AUDIO", + "link": 10 + }, + { + "name": "speaker_3_voice", + "shape": 7, + "type": "AUDIO", + "link": 6 + }, + { + "name": "speaker_4_voice", + "shape": 7, + "type": "AUDIO", + "link": 8 + } + ], + "outputs": [ + { + "name": "AUDIO", + "type": "AUDIO", + "links": [ + 2, + 4 + ] + } + ], + "properties": { + "cnr_id": "ComfyUI-VibeVoice", + "ver": "d04665d073dd9a587e682422deba5927e31ccac4", + "Node name for S&R": "VibeVoiceTTS" + }, + "widgets_values": [ + "VibeVoice-1.5B", + "Speaker 1: Hey, how are you?\nSpeaker 2: I'm fine, and you?", + false, + "sdpa", + 1.3, + 10, + 1114540226990889, + "randomize", + true, + 0.95, + 0.95, + 0, + false + ], + "color": "#232", + "bgcolor": "#353" + }, + { + "id": 14, + "type": "Note", + "pos": [ + -30.856473922729492, + 251.25491333007812 + ], + "size": [ + 393.2325744628906, + 708.2105102539062 + ], + "flags": {}, + "order": 4, + "mode": 0, + "inputs": [], + "outputs": [], + "properties": {}, + "widgets_values": [ + "# VibeVoice TTS\n\nThis workflow is based on the following git repository:\nhttps://github.com/wildminder/ComfyUI-VibeVoice\n\n# License\n\nAs stated on the git repository: This project is distributed under the MIT License. See LICENSE.txt (in the git repository above) for more information. The VibeVoice model and its components are subject to the licenses provided by Microsoft. Please use responsibly.\n\n# Enable more speakers\n\nClick the LoadAudio node and press Ctrl + B to enable the other speaker nodes.\n" + ], + "color": "#232", + "bgcolor": "#353" + } + ], + "links": [ + [ + 1, + 2, + 0, + 1, + 0, + "AUDIO" + ], + [ + 2, + 1, + 0, + 3, + 0, + "AUDIO" + ], + [ + 4, + 1, + 0, + 5, + 0, + "AUDIO" + ], + [ + 6, + 7, + 0, + 1, + 2, + "AUDIO" + ], + [ + 8, + 8, + 0, + 1, + 3, + "AUDIO" + ], + [ + 10, + 10, + 0, + 1, + 1, + "AUDIO" + ] + ], + "groups": [], + "config": {}, + "extra": { + "ds": { + "scale": 0.6303940863128474, + "offset": [ + 244.63564633654366, + -175.70438973601776 + ] + }, + "frontendVersion": "1.24.4", + "VHS_latentpreview": false, + "VHS_latentpreviewrate": 0, + "VHS_MetadataImage": true, + "VHS_KeepIntermediate": true + }, + "version": 0.4 +} \ No newline at end of file