跳转至内容
  • 版块
  • 最新
  • 标签
  • 热门
  • 用户
  • 群组
皮肤
  • 浅色
  • Brite
  • Cerulean
  • Cosmo
  • Flatly
  • Journal
  • Litera
  • Lumen
  • Lux
  • Materia
  • Minty
  • Morph
  • Pulse
  • Sandstone
  • Simplex
  • Sketchy
  • Spacelab
  • United
  • Yeti
  • Zephyr
  • 深色
  • Cyborg
  • Darkly
  • Quartz
  • Slate
  • Solar
  • Superhero
  • Vapor

  • 默认(LCZ-Blue)
  • 不使用皮肤
  • LCZ-Green
  • LCZ-Blue
  • LCZ-Black
折叠
品牌标识

抡锤者

首页 版块 标签
硬件
装备秀
AI 广场 App
  1. 主页
  2. 版块
  3. AI音视频画图
  4. [H3剑心比赛] RTX 5080 16GB 单卡|MiniMax H3 3D / 2D|灰盒深度控动作+STEP600 8 步

[H3剑心比赛] RTX 5080 16GB 单卡|MiniMax H3 3D / 2D|灰盒深度控动作+STEP600 8 步

已定时 已固定 已锁定 已移动 优秀 AI音视频画图
rtx5080minimax视频生成
3 帖子 3 发布者 165 浏览
  • 从旧到新
  • 从新到旧
  • 最多赞同
回复
  • 在新帖中回复
登录后回复
此主题已被删除。只有拥有主题管理权限的用户可以查看。
  • ゆうみんゆ
    ゆうみんゆ
    ゆうみん
    体面乡绅 硕果累累
    编写于 最后由 编辑
    #1

    3D CGI 成片:https://youtu.be/lXGIlkPRJN4
    传统 2D 动画成片:https://youtu.be/GaWO4hwy53A

    成片 = H3 原样画面(1344×768,取前 288 帧=12.00 秒)+后期配音、字幕、RTX 超分到 2560×1440。画面本身没有补帧、裁切、重绘。和别人不一样的地方在下面「说明」里照实写了。

    关键帧(3D)
    关键帧_3D.jpg

    关键帧(2D)
    关键帧_2D.jpg

    硬件

    • 显卡:RTX 5080 16GB(驱动 616.92,报告 16,303 MiB)
    • CPU:Ryzen 9 9900X|内存:128GB|SSD:Crucial P310 2TB
    • 系统:Windows 11 Pro(build 26300)|ComfyUI 0.44.0 便携版|torch 2.14.1+cu130|Python 3.13

    工作流(两条完全相同,只换画风和参考图)

    • 底模:minimax_h3_hybrid_fl2va_ref2va_b25-49-int8
    • 加速 LoRA:larryvrh minimax_h3_turbo_v4_step600_ema,强度 1.0(节点 MiniMaxH3TurboLoRA+MiniMaxH3TurboSampler,来自 ComfyUI-MiniMax-H3-Turbo)
    • 采样:8 步,simple,BasicGuider,种子 1(两条同一个)
    • 文本编码器 qwen3vl_32b_minimax_h3_nvfp4_awq;视频 VAE fp16+VAEDecodeTiled;音频 VAE fp32
    • 参考图:ComfyUI 自带 MiniMax H3 Reference to Video
      • 3D:剑心手办+志志雄 3 张手办(共 4 张)
      • 2D:剑心 2 张动画图+志志雄 3 张手办(共 5 张)
    • 控制:Fun ControlNet Union 2.0(剪枝 int8),自己做的灰盒深度视频,强度 0.6
    • 1344×768,生成 294 帧(H3 要 17n+5 帧),导出前 288 帧=24fps 12.00 秒
    • 提示词:比赛原文逐字不改,只把 Style 二选一
    • 工作流 JSON(API 格式)在文末

    加速手段

    • int8 剪枝底模+STEP600 加速 LoRA 8 步
    • comfy kitchen attention(全量注意力,不用稀疏)
    • 16G 显存跑 1344×768×294 帧+控制网:--disable-dynamic-vram --reserve-vram 9,再加两个分块补丁(MLP 和 LoRA 的线性层按 16384 个 token 分块算,数学完全一样,只是省显存)

    数据

    版本 Total Time ComfyUI 执行 8 步采样 显存峰值 内存峰值 显卡功耗 最高温度
    3D 762.4 s 12:39 10:17(77.2 s/步) 15,803 MiB 74.3 GB 344 W 79°C
    2D 719.8 s 11:57 10:23(78.0 s/步) 15,779 MiB 69.0 GB 346 W 79°C

    Total Time=提交工作流到 MP4 写完(含读模型、文本和参考图编码、采样、视频和声音解码)。3D 是 ComfyUI 启动后第一条,所以多了读模型的时间。

    运行日志(3D)
    日志_3D.png

    运行日志(2D)
    日志_2D.png

    说明:和别人不一样的地方(照实写)

    1. 用了灰盒深度控制。 自己用 Python 画了方块人灰盒(红=剑心、紫=志志雄),排好 12 秒的走位和机位,导出深度视频喂给控制网。强度 1.0 时两个人会认反(深度图没颜色),降到 0.6 才对。提示词没改。

    灰盒预览.jpg
    2. 声音全是后期配的。 H3 自带的声音不用;日语喊声用 Gemini TTS,撞击声用 Stable Audio 3 Medium 做,再按画面逐帧对上撞剑时刻;中文字幕。
    3. 放大。 成片用 NVIDIA RTX 超分(ComfyUI 的 RTXVideoSuperResolution 节点)放大到 2560×1440,加了一点对比和锐化。

    过程:快动作「划痕」是怎么解决的

    一开始用 minimax_h3_ref2v_turbo_4step_v0.1 走 4 步,快动作时人物身上一道道划痕,像油画笔触(发过求助帖 https://lcz.me/topic/3092 )。
    按 @perter 的建议换成 larryvrh 的 STEP600、走 8 步,划痕基本没了。同一个 4 步 LoRA 加到 8 步几乎没用,问题出在那个 LoRA 本身。谢谢 perter!
    左=旧版(4 步 LoRA),右=这次(STEP600 8 步):
    旧版vsSTEP600_3D.jpg

    自评

    • 好的:动作顺序按灰盒走(冲刺 → 横扫 → 相撞 → 对砍 → 落地收势),快动作清楚了;3D 手办质感和火光比较足。
    • 不够的:2D 版剑心头发在 3~8.5 秒偏金黄;2D 有两处灰色冲击波环;3D 撞剑白光被背景的火盖住、不够亮;加了灰盒后每条要 12 分钟左右,比不加慢。

    复现

    • 参考图:比赛原帖 6 张(3D 用 4 张、2D 用 5 张,见上)
    • 灰盒深度视频原件 2.6MB,超过论坛 2MB 上限;需要的话私信我
    • 工作流 JSON(3D,API 格式):
    {
     "unet": {
      "class_type": "UNETLoader",
      "inputs": {
       "unet_name": "minimax_h3_hybrid_fl2va_ref2va_b25-49-int8.safetensors",
       "weight_dtype": "default"
      }
     },
     "clip": {
      "class_type": "CLIPLoader",
      "inputs": {
       "clip_name": "qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors",
       "type": "minimax",
       "device": "default"
      }
     },
     "vae": {
      "class_type": "VAELoader",
      "inputs": {
       "vae_name": "minimax_h3_video_vae_fp16.safetensors"
      }
     },
     "cond": {
      "class_type": "MiniMaxH3ReferenceToVideo",
      "inputs": {
       "clip": [
        "clip",
        0
       ],
       "vae": [
        "vae",
        0
       ],
       "audio_vae": [
        "avae",
        0
       ],
       "prompt": "(Style: 3D CGI animation) Masterpiece, high quality, cinematic lighting. Subject: A fierce battle sequence featuring two characters based on the reference images—a red-clad swordsman (Jianxin) and a purple-robed warrior (Zhizhixiong). Action: The camera begins with a dynamic low-angle tracking shot. The swordsman in red dashes forward with a long sword, leaving a glowing trail, while the warrior in purple stands firm and executes a sweeping counter-attack with a flaming effect. They clash in the center of the frame, creating a shockwave of energy and sparks. The camera rapidly orbits around them as the fiery battle intensifies, concluding with the red-clad swordsman landing gracefully while the purple warrior points his weapon forward, both facing the camera. Environment: Dramatic background with embers floating in the air, high contrast lighting.",
       "width": 1344,
       "height": 768,
       "length": 294,
       "ref_image_size": "match",
       "ref_images.ref_image_0": [
        "rimg0",
        0
       ],
       "ref_images.ref_image_1": [
        "rimg1",
        0
       ],
       "ref_images.ref_image_2": [
        "rimg2",
        0
       ],
       "ref_images.ref_image_3": [
        "rimg3",
        0
       ]
      }
     },
     "noise": {
      "class_type": "RandomNoise",
      "inputs": {
       "noise_seed": 1
      }
     },
     "dec": {
      "class_type": "VAEDecodeTiled",
      "inputs": {
       "samples": [
        "ksa",
        0
       ],
       "vae": [
        "vae",
        0
       ],
       "tile_size": 512,
       "overlap": 64,
       "temporal_size": 64,
       "temporal_overlap": 8
      }
     },
     "save": {
      "class_type": "SaveVideo",
      "inputs": {
       "video": [
        "mk",
        0
       ],
       "filename_prefix": "h3test/K3D_s600",
       "format": "auto",
       "codec": "auto"
      }
     },
     "lora": {
      "class_type": "MiniMaxH3TurboLoRA",
      "inputs": {
       "model": [
        "unet",
        0
       ],
       "lora_name": "minimax_h3_turbo_v4_step600_ema.safetensors",
       "strength": 1.0,
       "low_vram": false
      }
     },
     "attention": {
      "class_type": "ModelAttentionBackend",
      "inputs": {
       "model": [
        "lora",
        0
       ],
       "attention": "comfy kitchen attention"
      }
     },
     "sampler": {
      "class_type": "MiniMaxH3TurboSampler",
      "inputs": {}
     },
     "patch": {
      "class_type": "ModelPatchLoader",
      "inputs": {
       "name": "minimax_h3_fun_controlnet_union_2.0_pruned_int8_convrot.safetensors"
      }
     },
     "lv": {
      "class_type": "LoadVideo",
      "inputs": {
       "file": "kenshin_ctrl_v3b_depth_1344x768_294f.mp4"
      }
     },
     "comp": {
      "class_type": "GetVideoComponents",
      "inputs": {
       "video": [
        "lv",
        0
       ]
      }
     },
     "cn": {
      "class_type": "MiniMaxH3FunControlNetApply",
      "inputs": {
       "model": [
        "attention",
        0
       ],
       "model_patch": [
        "patch",
        0
       ],
       "vae": [
        "vae",
        0
       ],
       "strength": 0.6,
       "start_percent": 0.0,
       "end_percent": 1.0,
       "control_video": [
        "comp",
        0
       ]
      }
     },
     "avae": {
      "class_type": "VAELoader",
      "inputs": {
       "vae_name": "minimax_h3_audio_vae_fp32.safetensors"
      }
     },
     "rimg0": {
      "class_type": "LoadImage",
      "inputs": {
       "image": "kenshin_ref3.jpg"
      }
     },
     "rimg1": {
      "class_type": "LoadImage",
      "inputs": {
       "image": "kenshin_ref4.jpg"
      }
     },
     "rimg2": {
      "class_type": "LoadImage",
      "inputs": {
       "image": "kenshin_ref5.jpg"
      }
     },
     "rimg3": {
      "class_type": "LoadImage",
      "inputs": {
       "image": "kenshin_ref6.jpg"
      }
     },
     "sched": {
      "class_type": "BasicScheduler",
      "inputs": {
       "model": [
        "cn",
        0
       ],
       "scheduler": "simple",
       "steps": 8,
       "denoise": 1.0
      }
     },
     "guider": {
      "class_type": "BasicGuider",
      "inputs": {
       "model": [
        "cn",
        0
       ],
       "conditioning": [
        "cond",
        0
       ]
      }
     },
     "ksa": {
      "class_type": "SamplerCustomAdvanced",
      "inputs": {
       "noise": [
        "noise",
        0
       ],
       "guider": [
        "guider",
        0
       ],
       "sampler": [
        "sampler",
        0
       ],
       "sigmas": [
        "sched",
        0
       ],
       "latent_image": [
        "cond",
        1
       ]
      }
     },
     "adec": {
      "class_type": "VAEDecodeAudio",
      "inputs": {
       "samples": [
        "ksa",
        0
       ],
       "vae": [
        "avae",
        0
       ]
      }
     },
     "mk": {
      "class_type": "CreateVideo",
      "inputs": {
       "images": [
        "dec",
        0
       ],
       "fps": 24.0,
       "audio": [
        "adec",
        0
       ]
      }
     }
    }
    
    • 工作流 JSON(2D,API 格式):
    {
     "unet": {
      "class_type": "UNETLoader",
      "inputs": {
       "unet_name": "minimax_h3_hybrid_fl2va_ref2va_b25-49-int8.safetensors",
       "weight_dtype": "default"
      }
     },
     "clip": {
      "class_type": "CLIPLoader",
      "inputs": {
       "clip_name": "qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors",
       "type": "minimax",
       "device": "default"
      }
     },
     "vae": {
      "class_type": "VAELoader",
      "inputs": {
       "vae_name": "minimax_h3_video_vae_fp16.safetensors"
      }
     },
     "cond": {
      "class_type": "MiniMaxH3ReferenceToVideo",
      "inputs": {
       "clip": [
        "clip",
        0
       ],
       "vae": [
        "vae",
        0
       ],
       "audio_vae": [
        "avae",
        0
       ],
       "prompt": "(Style: 2D traditional anime) Masterpiece, high quality, cinematic lighting. Subject: A fierce battle sequence featuring two characters based on the reference images—a red-clad swordsman (Jianxin) and a purple-robed warrior (Zhizhixiong). Action: The camera begins with a dynamic low-angle tracking shot. The swordsman in red dashes forward with a long sword, leaving a glowing trail, while the warrior in purple stands firm and executes a sweeping counter-attack with a flaming effect. They clash in the center of the frame, creating a shockwave of energy and sparks. The camera rapidly orbits around them as the fiery battle intensifies, concluding with the red-clad swordsman landing gracefully while the purple warrior points his weapon forward, both facing the camera. Environment: Dramatic background with embers floating in the air, high contrast lighting.",
       "width": 1344,
       "height": 768,
       "length": 294,
       "ref_image_size": "match",
       "ref_images.ref_image_0": [
        "rimg0",
        0
       ],
       "ref_images.ref_image_1": [
        "rimg1",
        0
       ],
       "ref_images.ref_image_2": [
        "rimg2",
        0
       ],
       "ref_images.ref_image_3": [
        "rimg3",
        0
       ],
       "ref_images.ref_image_4": [
        "rimg4",
        0
       ]
      }
     },
     "noise": {
      "class_type": "RandomNoise",
      "inputs": {
       "noise_seed": 1
      }
     },
     "dec": {
      "class_type": "VAEDecodeTiled",
      "inputs": {
       "samples": [
        "ksa",
        0
       ],
       "vae": [
        "vae",
        0
       ],
       "tile_size": 512,
       "overlap": 64,
       "temporal_size": 64,
       "temporal_overlap": 8
      }
     },
     "save": {
      "class_type": "SaveVideo",
      "inputs": {
       "video": [
        "mk",
        0
       ],
       "filename_prefix": "h3test/K2D_s600",
       "format": "auto",
       "codec": "auto"
      }
     },
     "lora": {
      "class_type": "MiniMaxH3TurboLoRA",
      "inputs": {
       "model": [
        "unet",
        0
       ],
       "lora_name": "minimax_h3_turbo_v4_step600_ema.safetensors",
       "strength": 1.0,
       "low_vram": false
      }
     },
     "attention": {
      "class_type": "ModelAttentionBackend",
      "inputs": {
       "model": [
        "lora",
        0
       ],
       "attention": "comfy kitchen attention"
      }
     },
     "sampler": {
      "class_type": "MiniMaxH3TurboSampler",
      "inputs": {}
     },
     "patch": {
      "class_type": "ModelPatchLoader",
      "inputs": {
       "name": "minimax_h3_fun_controlnet_union_2.0_pruned_int8_convrot.safetensors"
      }
     },
     "lv": {
      "class_type": "LoadVideo",
      "inputs": {
       "file": "kenshin_ctrl_v3b_depth_1344x768_294f.mp4"
      }
     },
     "comp": {
      "class_type": "GetVideoComponents",
      "inputs": {
       "video": [
        "lv",
        0
       ]
      }
     },
     "cn": {
      "class_type": "MiniMaxH3FunControlNetApply",
      "inputs": {
       "model": [
        "attention",
        0
       ],
       "model_patch": [
        "patch",
        0
       ],
       "vae": [
        "vae",
        0
       ],
       "strength": 0.6,
       "start_percent": 0.0,
       "end_percent": 1.0,
       "control_video": [
        "comp",
        0
       ]
      }
     },
     "avae": {
      "class_type": "VAELoader",
      "inputs": {
       "vae_name": "minimax_h3_audio_vae_fp32.safetensors"
      }
     },
     "rimg0": {
      "class_type": "LoadImage",
      "inputs": {
       "image": "kenshin_ref1.jpg"
      }
     },
     "rimg1": {
      "class_type": "LoadImage",
      "inputs": {
       "image": "kenshin_ref2.jpg"
      }
     },
     "rimg2": {
      "class_type": "LoadImage",
      "inputs": {
       "image": "kenshin_ref4.jpg"
      }
     },
     "rimg3": {
      "class_type": "LoadImage",
      "inputs": {
       "image": "kenshin_ref5.jpg"
      }
     },
     "rimg4": {
      "class_type": "LoadImage",
      "inputs": {
       "image": "kenshin_ref6.jpg"
      }
     },
     "sched": {
      "class_type": "BasicScheduler",
      "inputs": {
       "model": [
        "cn",
        0
       ],
       "scheduler": "simple",
       "steps": 8,
       "denoise": 1.0
      }
     },
     "guider": {
      "class_type": "BasicGuider",
      "inputs": {
       "model": [
        "cn",
        0
       ],
       "conditioning": [
        "cond",
        0
       ]
      }
     },
     "ksa": {
      "class_type": "SamplerCustomAdvanced",
      "inputs": {
       "noise": [
        "noise",
        0
       ],
       "guider": [
        "guider",
        0
       ],
       "sampler": [
        "sampler",
        0
       ],
       "sigmas": [
        "sched",
        0
       ],
       "latent_image": [
        "cond",
        1
       ]
      }
     },
     "adec": {
      "class_type": "VAEDecodeAudio",
      "inputs": {
       "samples": [
        "ksa",
        0
       ],
       "vae": [
        "avae",
        0
       ]
      }
     },
     "mk": {
      "class_type": "CreateVideo",
      "inputs": {
       "images": [
        "dec",
        0
       ],
       "fps": 24.0,
       "audio": [
        "adec",
        0
       ]
      }
     }
    }
    

    youtube.com/channel/UCN2nQbk-8CJXHNMr4yWNt9Q/
    www.youtube.com/@AIdopagaki

    1 条回复 最后回复
    0
    • ,ゆうみんゆ ゆうみん 引用了 此主题
    • XiaoteX
      XiaoteX
      Xiaote
      易有太极
      编写于 最后由 编辑
      #2

      ✅ 版主审定:本帖设为优秀,作者 +3 积分奖励,继续加油!

      奖励凭证:优秀 +3 分 · 编号 e3104-4790-1

      老特的AI助手,DeepSeek Flash驱动,没回你是因为被限速了~除了管理员不能私信我,别浪费时间~尝试多了被封号就不好了~

      1 条回复 最后回复
      0
      • terryT
        terryT
        terry
        易有太极
        编写于 最后由 编辑
        #3

        过审,20积分比赛结束后统一结算。画面不错,你的声音约束下,不过也能接受。

        油管:https://www.youtube.com/@抡锤者

        1 条回复 最后回复
        0

        你好!看起来您对这段对话很感兴趣,但您还没有一个账号。

        厌倦了每次访问都刷到同样的帖子?您注册账号后,您下次访问时都将自动回到上次浏览的位置,并可选择接收新回复的通知(通过电子邮件或推送通知)。您还可以收藏帖子、为帖子点赞,以此向其他社区成员表达您的感谢。

        有了你的建议,这篇帖子会更精彩哦 💗

        注册 登录
        回复
        • 在新帖中回复
        登录后回复
        • 从旧到新
        • 从新到旧
        • 最多赞同


        • 登录

        • 登录或注册以进行搜索。
        • 第一个帖子
          最后一个帖子
        0
        • 版块
        • 最新
        • 标签
        • 热门
        • 用户
        • 群组