diff --git a/examples/cogvideox_5b_Tora_I2V_testing_01.json b/examples/cogvideox_5b_Tora_I2V_testing_01.json index 39f1ed1..604e0fe 100644 --- a/examples/cogvideox_5b_Tora_I2V_testing_01.json +++ b/examples/cogvideox_5b_Tora_I2V_testing_01.json @@ -1,6 +1,6 @@ { - "last_node_id": 75, - "last_link_id": 176, + "last_node_id": 77, + "last_link_id": 186, "nodes": [ { "id": 20, @@ -63,7 +63,7 @@ "name": "conditioning", "type": "CONDITIONING", "links": [ - 123 + 179 ], "slot_index": 0, "shape": 3 @@ -87,7 +87,7 @@ }, "size": { "0": 313.4619445800781, - "1": 270 + "1": 286 }, "flags": {}, "order": 8, @@ -162,7 +162,8 @@ 1, [ 1 - ] + ], + 1 ] }, { @@ -229,7 +230,7 @@ "hidden": false, "paused": false, "params": { - "filename": "CogVideoX-Tora-trajectory_00011.mp4", + "filename": "CogVideoX-Tora-trajectory_00003.mp4", "subfolder": "", "type": "temp", "format": "video/h264-mp4", @@ -257,12 +258,12 @@ { "name": "pipeline", "type": "COGVIDEOPIPE", - "link": 128 + "link": 184 }, { "name": "samples", "type": "LATENT", - "link": 127 + "link": 185 } ], "outputs": [ @@ -288,125 +289,6 @@ true ] }, - { - "id": 57, - "type": "CogVideoSampler", - "pos": { - "0": 1138, - "1": 150 - }, - "size": { - "0": 399.8780822753906, - "1": 390 - }, - "flags": {}, - "order": 13, - "mode": 0, - "inputs": [ - { - "name": "pipeline", - "type": "COGVIDEOPIPE", - "link": 121 - }, - { - "name": "positive", - "type": "CONDITIONING", - "link": 122 - }, - { - "name": "negative", - "type": "CONDITIONING", - "link": 123 - }, - { - "name": "samples", - "type": "LATENT", - "link": null, - "shape": 7 - }, - { - "name": "image_cond_latents", - "type": "LATENT", - "link": 162, - "shape": 7 - }, - { - "name": "context_options", - "type": "COGCONTEXT", - "link": null, - "shape": 7 - }, - { - "name": "controlnet", - "type": "COGVIDECONTROLNET", - "link": null, - "shape": 7 - }, - { - "name": "tora_trajectory", - "type": "TORAFEATURES", - "link": 173, - "shape": 7 - }, - { - "name": "num_frames", - "type": "INT", - "link": 157, - "widget": { - "name": "num_frames" - } - }, - { - "name": "height", - "type": "INT", - "link": 151, - "widget": { - "name": "height" - } - }, - { - "name": "width", - "type": "INT", - "link": 152, - "widget": { - "name": "width" - } - } - ], - "outputs": [ - { - "name": "cogvideo_pipe", - "type": "COGVIDEOPIPE", - "links": [ - 128 - ], - "slot_index": 0, - "shape": 3 - }, - { - "name": "samples", - "type": "LATENT", - "links": [ - 127 - ], - "shape": 3 - } - ], - "properties": { - "Node name for S&R": "CogVideoSampler" - }, - "widgets_values": [ - 480, - 720, - 49, - 32, - 6, - 65334758276105, - "fixed", - "CogVideoXDPMScheduler", - 1 - ] - }, { "id": 71, "type": "CogVideoImageEncode", @@ -444,7 +326,7 @@ "name": "samples", "type": "LATENT", "links": [ - 162 + 180 ], "slot_index": 0 } @@ -457,72 +339,6 @@ false ] }, - { - "id": 67, - "type": "GetMaskSizeAndCount", - "pos": { - "0": 763, - "1": 772 - }, - "size": { - "0": 264.5999755859375, - "1": 86 - }, - "flags": { - "collapsed": true - }, - "order": 7, - "mode": 0, - "inputs": [ - { - "name": "mask", - "type": "MASK", - "link": 146 - } - ], - "outputs": [ - { - "name": "mask", - "type": "MASK", - "links": null - }, - { - "name": "width", - "type": "INT", - "links": [ - 149, - 152, - 168, - 171 - ], - "slot_index": 1 - }, - { - "name": "height", - "type": "INT", - "links": [ - 150, - 151, - 169, - 172 - ], - "slot_index": 2 - }, - { - "name": "count", - "type": "INT", - "links": [ - 157, - 170 - ], - "slot_index": 3 - } - ], - "properties": { - "Node name for S&R": "GetMaskSizeAndCount" - }, - "widgets_values": [] - }, { "id": 72, "type": "LoadImage", @@ -705,7 +521,14 @@ "flags": {}, "order": 2, "mode": 0, - "inputs": [], + "inputs": [ + { + "name": "bg_image", + "type": "IMAGE", + "link": null, + "shape": 7 + } + ], "outputs": [ { "name": "mask", @@ -758,6 +581,7 @@ 0, 1, null, + null, null ] }, @@ -787,7 +611,7 @@ "name": "conditioning", "type": "CONDITIONING", "links": [ - 122 + 178 ], "slot_index": 0, "shape": 3 @@ -811,7 +635,7 @@ }, "size": [ 1131.619140625, - 1065.0794270833335 + 310 ], "flags": {}, "order": 16, @@ -914,10 +738,10 @@ "0": 1060, "1": 670 }, - "size": [ - 335.1993359916705, - 206 - ], + "size": { + "0": 335.1993408203125, + "1": 230 + }, "flags": {}, "order": 10, "mode": 0, @@ -970,7 +794,7 @@ "name": "tora_trajectory", "type": "TORAFEATURES", "links": [ - 173 + 181 ] }, { @@ -989,7 +813,8 @@ 49, 1, 0, - 0.1 + 0.1, + false ] }, { @@ -1031,9 +856,9 @@ "name": "cogvideo_pipe", "type": "COGVIDEOPIPE", "links": [ - 121, 164, - 174 + 174, + 177 ], "slot_index": 0, "shape": 3 @@ -1049,6 +874,194 @@ "disabled", false ] + }, + { + "id": 77, + "type": "CogVideoSampler", + "pos": { + "0": 1138, + "1": 150 + }, + "size": [ + 405.5999755859375, + 410 + ], + "flags": {}, + "order": 13, + "mode": 0, + "inputs": [ + { + "name": "pipeline", + "type": "COGVIDEOPIPE", + "link": 177 + }, + { + "name": "positive", + "type": "CONDITIONING", + "link": 178 + }, + { + "name": "negative", + "type": "CONDITIONING", + "link": 179 + }, + { + "name": "samples", + "type": "LATENT", + "link": null, + "shape": 7 + }, + { + "name": "image_cond_latents", + "type": "LATENT", + "link": 180, + "shape": 7 + }, + { + "name": "context_options", + "type": "COGCONTEXT", + "link": null, + "shape": 7 + }, + { + "name": "controlnet", + "type": "COGVIDECONTROLNET", + "link": null, + "shape": 7 + }, + { + "name": "tora_trajectory", + "type": "TORAFEATURES", + "link": 181, + "shape": 7 + }, + { + "name": "fastercache", + "type": "FASTERCACHEARGS", + "link": null, + "shape": 7 + }, + { + "name": "num_frames", + "type": "INT", + "link": 186, + "widget": { + "name": "num_frames" + } + }, + { + "name": "height", + "type": "INT", + "link": 182, + "widget": { + "name": "height" + } + }, + { + "name": "width", + "type": "INT", + "link": 183, + "widget": { + "name": "width" + } + } + ], + "outputs": [ + { + "name": "cogvideo_pipe", + "type": "COGVIDEOPIPE", + "links": [ + 184 + ] + }, + { + "name": "samples", + "type": "LATENT", + "links": [ + 185 + ] + } + ], + "properties": { + "Node name for S&R": "CogVideoSampler" + }, + "widgets_values": [ + 480, + 720, + 49, + 32, + 6, + 65334758276105, + "fixed", + "CogVideoXDPMScheduler", + 1 + ] + }, + { + "id": 67, + "type": "GetMaskSizeAndCount", + "pos": { + "0": 763, + "1": 772 + }, + "size": { + "0": 264.5999755859375, + "1": 86 + }, + "flags": { + "collapsed": true + }, + "order": 7, + "mode": 0, + "inputs": [ + { + "name": "mask", + "type": "MASK", + "link": 146 + } + ], + "outputs": [ + { + "name": "mask", + "type": "MASK", + "links": null + }, + { + "name": "720 width", + "type": "INT", + "links": [ + 149, + 168, + 171, + 183 + ], + "slot_index": 1 + }, + { + "name": "480 height", + "type": "INT", + "links": [ + 150, + 169, + 172, + 182 + ], + "slot_index": 2 + }, + { + "name": "49 count", + "type": "INT", + "links": [ + 170, + 186 + ], + "slot_index": 3 + } + ], + "properties": { + "Node name for S&R": "GetMaskSizeAndCount" + }, + "widgets_values": [] } ], "links": [ @@ -1068,46 +1081,6 @@ 0, "CLIP" ], - [ - 121, - 1, - 0, - 57, - 0, - "COGVIDEOPIPE" - ], - [ - 122, - 30, - 0, - 57, - 1, - "CONDITIONING" - ], - [ - 123, - 31, - 0, - 57, - 2, - "CONDITIONING" - ], - [ - 127, - 57, - 1, - 56, - 1, - "LATENT" - ], - [ - 128, - 57, - 0, - 56, - 0, - "COGVIDEOPIPE" - ], [ 142, 65, @@ -1148,22 +1121,6 @@ 3, "INT" ], - [ - 151, - 67, - 2, - 57, - 9, - "INT" - ], - [ - 152, - 67, - 1, - 57, - 10, - "INT" - ], [ 153, 65, @@ -1196,22 +1153,6 @@ 0, "IMAGE" ], - [ - 157, - 67, - 3, - 57, - 8, - "INT" - ], - [ - 162, - 71, - 0, - 57, - 4, - "LATENT" - ], [ 164, 1, @@ -1276,14 +1217,6 @@ 5, "INT" ], - [ - 173, - 74, - 0, - 57, - 7, - "TORAFEATURES" - ], [ 174, 1, @@ -1307,16 +1240,96 @@ 74, 2, "STRING" + ], + [ + 177, + 1, + 0, + 77, + 0, + "COGVIDEOPIPE" + ], + [ + 178, + 30, + 0, + 77, + 1, + "CONDITIONING" + ], + [ + 179, + 31, + 0, + 77, + 2, + "CONDITIONING" + ], + [ + 180, + 71, + 0, + 77, + 4, + "LATENT" + ], + [ + 181, + 74, + 0, + 77, + 7, + "TORAFEATURES" + ], + [ + 182, + 67, + 2, + 77, + 10, + "INT" + ], + [ + 183, + 67, + 1, + 77, + 11, + "INT" + ], + [ + 184, + 77, + 0, + 56, + 0, + "COGVIDEOPIPE" + ], + [ + 185, + 77, + 1, + 56, + 1, + "LATENT" + ], + [ + 186, + 67, + 3, + 77, + 9, + "INT" ] ], "groups": [], "config": {}, "extra": { "ds": { - "scale": 0.5730855330117661, + "scale": 0.6303940863129501, "offset": [ - 931.6955110788911, - 94.46846235728464 + 136.37893073690276, + 136.88820468799213 ] } }, diff --git a/examples/cogvideox_5b_tora_trajectory_example_01.json b/examples/cogvideox_5b_tora_trajectory_example_01.json index bb74657..bad2570 100644 --- a/examples/cogvideox_5b_tora_trajectory_example_01.json +++ b/examples/cogvideox_5b_tora_trajectory_example_01.json @@ -1,6 +1,6 @@ { - "last_node_id": 70, - "last_link_id": 165, + "last_node_id": 72, + "last_link_id": 174, "nodes": [ { "id": 20, @@ -63,7 +63,7 @@ "name": "conditioning", "type": "CONDITIONING", "links": [ - 123 + 168 ], "slot_index": 0, "shape": 3 @@ -117,8 +117,8 @@ "name": "cogvideo_pipe", "type": "COGVIDEOPIPE", "links": [ - 121, - 159 + 159, + 166 ], "slot_index": 0, "shape": 3 @@ -144,7 +144,7 @@ }, "size": { "0": 313.4619445800781, - "1": 270 + "1": 286 }, "flags": {}, "order": 7, @@ -219,7 +219,8 @@ 1, [ 1 - ] + ], + 1 ] }, { @@ -240,12 +241,12 @@ { "name": "pipeline", "type": "COGVIDEOPIPE", - "link": 128 + "link": 172 }, { "name": "samples", "type": "LATENT", - "link": 127 + "link": 173 } ], "outputs": [ @@ -386,7 +387,7 @@ "hidden": false, "paused": false, "params": { - "filename": "CogVideoX-Tora-trajectory_00012.mp4", + "filename": "CogVideoX-Tora-trajectory_00001.mp4", "subfolder": "", "type": "temp", "format": "video/h264-mp4", @@ -422,7 +423,7 @@ "name": "conditioning", "type": "CONDITIONING", "links": [ - 122 + 167 ], "slot_index": 0, "shape": 3 @@ -451,7 +452,14 @@ "flags": {}, "order": 2, "mode": 0, - "inputs": [], + "inputs": [ + { + "name": "bg_image", + "type": "IMAGE", + "link": null, + "shape": 7 + } + ], "outputs": [ { "name": "mask", @@ -504,192 +512,10 @@ 0, 1, null, + null, null ] }, - { - "id": 57, - "type": "CogVideoSampler", - "pos": { - "0": 1138, - "1": 150 - }, - "size": { - "0": 399.8780822753906, - "1": 390 - }, - "flags": {}, - "order": 10, - "mode": 0, - "inputs": [ - { - "name": "pipeline", - "type": "COGVIDEOPIPE", - "link": 121 - }, - { - "name": "positive", - "type": "CONDITIONING", - "link": 122 - }, - { - "name": "negative", - "type": "CONDITIONING", - "link": 123 - }, - { - "name": "samples", - "type": "LATENT", - "link": null, - "shape": 7 - }, - { - "name": "image_cond_latents", - "type": "LATENT", - "link": null, - "shape": 7 - }, - { - "name": "context_options", - "type": "COGCONTEXT", - "link": null, - "shape": 7 - }, - { - "name": "controlnet", - "type": "COGVIDECONTROLNET", - "link": null, - "shape": 7 - }, - { - "name": "tora_trajectory", - "type": "TORAFEATURES", - "link": 163, - "shape": 7 - }, - { - "name": "num_frames", - "type": "INT", - "link": 157, - "widget": { - "name": "num_frames" - } - }, - { - "name": "height", - "type": "INT", - "link": 151, - "widget": { - "name": "height" - } - }, - { - "name": "width", - "type": "INT", - "link": 152, - "widget": { - "name": "width" - } - } - ], - "outputs": [ - { - "name": "cogvideo_pipe", - "type": "COGVIDEOPIPE", - "links": [ - 128 - ], - "slot_index": 0, - "shape": 3 - }, - { - "name": "samples", - "type": "LATENT", - "links": [ - 127 - ], - "shape": 3 - } - ], - "properties": { - "Node name for S&R": "CogVideoSampler" - }, - "widgets_values": [ - 480, - 720, - 49, - 32, - 6, - 65334758276105, - "fixed", - "CogVideoXDPMScheduler", - 1 - ] - }, - { - "id": 67, - "type": "GetMaskSizeAndCount", - "pos": { - "0": 763, - "1": 772 - }, - "size": { - "0": 264.5999755859375, - "1": 86 - }, - "flags": { - "collapsed": true - }, - "order": 6, - "mode": 0, - "inputs": [ - { - "name": "mask", - "type": "MASK", - "link": 146 - } - ], - "outputs": [ - { - "name": "mask", - "type": "MASK", - "links": null - }, - { - "name": "720 width", - "type": "INT", - "links": [ - 149, - 152, - 161 - ], - "slot_index": 1 - }, - { - "name": "480 height", - "type": "INT", - "links": [ - 150, - 151, - 162 - ], - "slot_index": 2 - }, - { - "name": "49 count", - "type": "INT", - "links": [ - 157, - 160 - ], - "slot_index": 3 - } - ], - "properties": { - "Node name for S&R": "GetMaskSizeAndCount" - }, - "widgets_values": [] - }, { "id": 44, "type": "VHS_VideoCombine", @@ -699,7 +525,7 @@ }, "size": [ 1131.619140625, - 1065.0794270833335 + 310 ], "flags": {}, "order": 13, @@ -802,10 +628,10 @@ "0": 1112, "1": 663 }, - "size": [ - 355.20001220703125, - 222 - ], + "size": { + "0": 355.20001220703125, + "1": 230 + }, "flags": {}, "order": 8, "mode": 0, @@ -858,7 +684,7 @@ "name": "tora_trajectory", "type": "TORAFEATURES", "links": [ - 163 + 169 ] }, { @@ -877,6 +703,193 @@ 49, 1, 0, + 1, + false + ] + }, + { + "id": 67, + "type": "GetMaskSizeAndCount", + "pos": { + "0": 763, + "1": 772 + }, + "size": { + "0": 264.5999755859375, + "1": 86 + }, + "flags": { + "collapsed": true + }, + "order": 6, + "mode": 0, + "inputs": [ + { + "name": "mask", + "type": "MASK", + "link": 146 + } + ], + "outputs": [ + { + "name": "mask", + "type": "MASK", + "links": null + }, + { + "name": "width", + "type": "INT", + "links": [ + 149, + 161, + 171 + ], + "slot_index": 1 + }, + { + "name": "height", + "type": "INT", + "links": [ + 150, + 162, + 170 + ], + "slot_index": 2 + }, + { + "name": "count", + "type": "INT", + "links": [ + 160, + 174 + ], + "slot_index": 3 + } + ], + "properties": { + "Node name for S&R": "GetMaskSizeAndCount" + }, + "widgets_values": [] + }, + { + "id": 72, + "type": "CogVideoSampler", + "pos": { + "0": 1138, + "1": 150 + }, + "size": { + "0": 405.5999755859375, + "1": 410 + }, + "flags": {}, + "order": 10, + "mode": 0, + "inputs": [ + { + "name": "pipeline", + "type": "COGVIDEOPIPE", + "link": 166 + }, + { + "name": "positive", + "type": "CONDITIONING", + "link": 167 + }, + { + "name": "negative", + "type": "CONDITIONING", + "link": 168 + }, + { + "name": "samples", + "type": "LATENT", + "link": null, + "shape": 7 + }, + { + "name": "image_cond_latents", + "type": "LATENT", + "link": null, + "shape": 7 + }, + { + "name": "context_options", + "type": "COGCONTEXT", + "link": null, + "shape": 7 + }, + { + "name": "controlnet", + "type": "COGVIDECONTROLNET", + "link": null, + "shape": 7 + }, + { + "name": "tora_trajectory", + "type": "TORAFEATURES", + "link": 169, + "shape": 7 + }, + { + "name": "fastercache", + "type": "FASTERCACHEARGS", + "link": null, + "shape": 7 + }, + { + "name": "num_frames", + "type": "INT", + "link": 174, + "widget": { + "name": "num_frames" + } + }, + { + "name": "height", + "type": "INT", + "link": 170, + "widget": { + "name": "height" + } + }, + { + "name": "width", + "type": "INT", + "link": 171, + "widget": { + "name": "width" + } + } + ], + "outputs": [ + { + "name": "cogvideo_pipe", + "type": "COGVIDEOPIPE", + "links": [ + 172 + ] + }, + { + "name": "samples", + "type": "LATENT", + "links": [ + 173 + ] + } + ], + "properties": { + "Node name for S&R": "CogVideoSampler" + }, + "widgets_values": [ + 480, + 720, + 49, + 32, + 6, + 65334758276105, + "fixed", + "CogVideoXDPMScheduler", 1 ] } @@ -898,46 +911,6 @@ 0, "CLIP" ], - [ - 121, - 1, - 0, - 57, - 0, - "COGVIDEOPIPE" - ], - [ - 122, - 30, - 0, - 57, - 1, - "CONDITIONING" - ], - [ - 123, - 31, - 0, - 57, - 2, - "CONDITIONING" - ], - [ - 127, - 57, - 1, - 56, - 1, - "LATENT" - ], - [ - 128, - 57, - 0, - 56, - 0, - "COGVIDEOPIPE" - ], [ 142, 65, @@ -978,22 +951,6 @@ 3, "INT" ], - [ - 151, - 67, - 2, - 57, - 9, - "INT" - ], - [ - 152, - 67, - 1, - 57, - 10, - "INT" - ], [ 153, 65, @@ -1026,14 +983,6 @@ 0, "IMAGE" ], - [ - 157, - 67, - 3, - 57, - 8, - "INT" - ], [ 159, 1, @@ -1066,14 +1015,6 @@ 5, "INT" ], - [ - 163, - 69, - 0, - 57, - 7, - "TORAFEATURES" - ], [ 164, 70, @@ -1089,16 +1030,88 @@ 69, 2, "STRING" + ], + [ + 166, + 1, + 0, + 72, + 0, + "COGVIDEOPIPE" + ], + [ + 167, + 30, + 0, + 72, + 1, + "CONDITIONING" + ], + [ + 168, + 31, + 0, + 72, + 2, + "CONDITIONING" + ], + [ + 169, + 69, + 0, + 72, + 7, + "TORAFEATURES" + ], + [ + 170, + 67, + 2, + 72, + 10, + "INT" + ], + [ + 171, + 67, + 1, + 72, + 11, + "INT" + ], + [ + 172, + 72, + 0, + 56, + 0, + "COGVIDEOPIPE" + ], + [ + 173, + 72, + 1, + 56, + 1, + "LATENT" + ], + [ + 174, + 67, + 3, + 72, + 9, + "INT" ] ], "groups": [], "config": {}, "extra": { "ds": { - "scale": 0.6303940863129383, + "scale": 0.7627768444386698, "offset": [ - 302.46497962544663, - 77.05651988573555 + -854.9458997028877, + 235.76201522394632 ] } },