docs: add H3 image and reference video examples

2026-11-01 17:13:04 +08:00
parent be22dbd5b0
commit 7fa989e89c
4 changed files with 2043 additions and 0 deletions
@@ -0,0 +1,182 @@
+{
+  "6": {
+    "class_type": "UNETLoader",
+    "inputs": {
+      "unet_name": "minimax_h3_fl2va_pruned_int8_convrot.safetensors",
+      "weight_dtype": "default"
+    }
+  },
+  "121": {
+    "class_type": "LoraLoaderModelOnly",
+    "inputs": {
+      "model": [
+        "6",
+        0
+      ],
+      "lora_name": "minimax_h3_fl2v_turbo_8step_v1.0_comfyui_bf16.safetensors",
+      "strength_model": 1.0
+    }
+  },
+  "13": {
+    "class_type": "CLIPLoader",
+    "inputs": {
+      "clip_name": "qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors",
+      "type": "minimax",
+      "device": "default"
+    }
+  },
+  "11": {
+    "class_type": "VAELoader",
+    "inputs": {
+      "vae_name": "minimax_h3_video_vae_fp16.safetensors"
+    }
+  },
+  "24": {
+    "class_type": "VAELoader",
+    "inputs": {
+      "vae_name": "minimax_h3_audio_vae_fp32.safetensors"
+    }
+  },
+  "104": {
+    "class_type": "MiniMaxH3ImageToVideo",
+    "inputs": {
+      "clip": [
+        "13",
+        0
+      ],
+      "vae": [
+        "11",
+        0
+      ],
+      "prompt": "For the target video, at 0.00 seconds into the target video, <Picture 1> (from [Shot 1]) is fully referenced.\n\nintegrated_multimodal_description: [Shot 1] Preserve the exact reference character identities, clothes, realistic face proportions, props and light direction. Live-action movie cinematography. Slow Push In, medium close-up of the SAME elderly Moses from the reference, exact face and rust-red mantle, only a slight head motion, looks toward unseen Pharaoh just camera right. At 00:00.600 he draws a restrained breath and holds his gaze. The SAME original elderly male Mandarin voice for Moses (S1): deep warm resonant chest baritone, about 70 years old, mildly textured breath but clear strong articulation, measured grounded film acting, standard Mandarin, never a young broadcaster, never an announcer or a chant. At 00:01.300 he says with quiet authority, with a firm pause after the first word and weight on the final word: <d>[Chinese] 让我的百姓走。</d> End by 00:04.800, then his lips close, eyes hold steady and jaw tightens, no shouting and no grin. No other person speaks, no narration or extra words.\n\noverall_soundscape: A huge quiet stone interior, soft torch crackle and faint distant air, only the specified elderly Mandarin sentence with natural lip synchronization.\n\nnon_diegetic_music: None.",
+      "width": 960,
+      "height": 544,
+      "length": 175,
+      "first_frame": [
+        "1",
+        0
+      ]
+    }
+  },
+  "15": {
+    "class_type": "RandomNoise",
+    "inputs": {
+      "noise_seed": 7331420710003001
+    }
+  },
+  "16": {
+    "class_type": "BasicGuider",
+    "inputs": {
+      "model": [
+        "121",
+        0
+      ],
+      "conditioning": [
+        "104",
+        0
+      ]
+    }
+  },
+  "17": {
+    "class_type": "KSamplerSelect",
+    "inputs": {
+      "sampler_name": "res_multistep"
+    }
+  },
+  "9": {
+    "class_type": "BasicScheduler",
+    "inputs": {
+      "model": [
+        "121",
+        0
+      ],
+      "scheduler": "simple",
+      "steps": 8,
+      "denoise": 1.0
+    }
+  },
+  "14": {
+    "class_type": "SamplerCustomAdvanced",
+    "inputs": {
+      "noise": [
+        "15",
+        0
+      ],
+      "guider": [
+        "16",
+        0
+      ],
+      "sampler": [
+        "17",
+        0
+      ],
+      "sigmas": [
+        "9",
+        0
+      ],
+      "latent_image": [
+        "104",
+        1
+      ]
+    }
+  },
+  "10": {
+    "class_type": "VAEDecode",
+    "inputs": {
+      "samples": [
+        "14",
+        0
+      ],
+      "vae": [
+        "11",
+        0
+      ]
+    }
+  },
+  "23": {
+    "class_type": "VAEDecodeAudio",
+    "inputs": {
+      "samples": [
+        "14",
+        0
+      ],
+      "vae": [
+        "24",
+        0
+      ]
+    }
+  },
+  "91": {
+    "class_type": "CreateVideo",
+    "inputs": {
+      "images": [
+        "10",
+        0
+      ],
+      "audio": [
+        "23",
+        0
+      ],
+      "fps": 24.0,
+      "bit_depth": 8
+    }
+  },
+  "92": {
+    "class_type": "SaveVideo",
+    "inputs": {
+      "video": [
+        "91",
+        0
+      ],
+      "filename_prefix": "exodus_20261030/S02_demand",
+      "format": "auto",
+      "codec": "auto"
+    }
+  },
+  "1": {
+    "class_type": "LoadImage",
+    "inputs": {
+      "image": "exodus_20261030/refs/K02_demand.png"
+    }
+  }
+}