想在本地用單一模型,就能高速生成高品質二次元動漫插圖,而且不需要額外下載 VAE、Text Encoder、LoRA 或複雜前處理?這篇文章要介紹的是一條極度精簡、但品質與速度兼顧的 ComfyUI 生圖工作流,核心模型為:Z-Image-Turbo-Anime(AIO)。
Z-Image-Turbo-Anime 是基於 Z-Image-Turbo 所打造的動漫向整合模型,透過多組專用 Anime LoRA 的合併訓練,在完全不犧牲 8-step 極速生成的前提下,大幅強化人物比例、臉部表情、髮絲細節與整體二次元風格一致性。
這條工作流的設計目標很單純: 用最少的節點,完成穩定、可重複、可量產的動漫插圖生成流程。你只需要專心寫好自然語言提示詞,其餘包含取樣策略、精度格式與編碼器整合,全都交給模型處理。
Z-Image-Turbo-Anime 是什麼?為什麼適合做圖片編輯?
Z-Image-Turbo-Anime 是以 Z-Image-Turbo 為基礎,針對「二次元插圖生成」重新調校的動漫專用模型。它並不是單純套用風格,而是透過多組專門為動漫比例與表現訓練的 LoRA 進行合併,直接改變模型在人物結構、臉部特徵與色彩輸出的行為。
它的幾個關鍵特性包括:
- 保留原生 8-step 超高速生成
- 人物臉型、眼睛比例更接近日系動漫
- 髮絲與服裝細節更銳利,顏色表現更乾淨
- All-in-One 封裝:VAE 與 Text Encoder 已整合
- 不需要額外下載或管理 Anime LoRA
在 ComfyUI 的實務使用上,它和傳統「Stable Diffusion + Anime LoRA」的流程差異非常明顯。過去如果你想要穩定產出動漫插圖,通常需要:
- 選模型 → 換 VAE
- 疊多個 Anime LoRA
- 調整 LoRA 權重
- 反覆試 sampler 與 step
而 Z-Image-Turbo-Anime 的做法是: 把這些決策全部提前收斂進模型本身,讓使用者只需要專注在「畫面描述」本身。
本次實測環境
硬體與測試條件:
- 顯示卡:RTX 5070 Ti 16GB VRAM
- 記憶體:64GB RAM
- 圖片解析度:800 × 1200
- 每批次生成張數:2 張(Batch size 2)
- 取樣步數:9 steps
- CFG Scale:1.0
- Sampler:Euler A(euler_ancestral)
- Scheduler:Beta
這組設定並不是為了追求單張圖片的極限畫質,而是刻意選擇官方建議、且最具代表性的取樣組合,用來驗證在中高解析度、Batch size 為 2 的情境下,模型是否仍能維持:
- 穩定的人物結構與動漫比例
- 一致的風格輸出,不易崩圖
- 足夠快的生成速度,適合高頻反覆嘗試
也就是說,這是一組偏向實務產能與穩定度驗證的設定,而非單純為了拉高單張畫質所做的極限調校。
環境配置與模型放置
請先完成 ComfyUI 的基本安裝並能正常啟動介面(教學點我)。接著,這套 AIO 模型的優勢就在於: 你只需要準備一個主要模型檔即可開始工作。
模型與路徑
Z-Image-Turbo-Anime 提供多種 AIO 精度版本[ 下載點],可依需求選擇其中一種:
- FP8-AIO(約 10GB):體積小、速度快,最適合日常使用
- FP16-AIO(約 20GB):相容性最佳,適合較舊或混合硬體
- BF16-AIO(約 20GB):精度最高,適合對細節要求較高的工作(👍 我使用這個模型)
📁 放置路徑:
ComfyUI/models/checkpoints/
與傳統流程不同的是: VAE 與 Text Encoder 已完整整合在 checkpoint 中,不需要額外下載或指定,模型載入後即可直接使用。
工作流(Z-Image-Turbo-Anime.json)
將下列內容複製到文字編輯器,儲存為 Z-Image-Turbo-Anime.json 後即可在 ComfyUI 匯入。
{"id":"3cd9e899-67c5-43fb-90bb-364bc5775bc4","revision":0,"last_node_id":57,"last_link_id":66,"nodes":[{"id":42,"type":"ConditioningZeroOut","pos":[657.9563923143061,727.5619449179278],"size":[204.134765625,26],"flags":{},"order":3,"mode":0,"inputs":[{"name":"conditioning","type":"CONDITIONING","link":36}],"outputs":[{"name":"CONDITIONING","type":"CONDITIONING","links":[42]}],"properties":{"Node name for S&R":"ConditioningZeroOut","cnr_id":"comfy-core","ver":"0.3.73","enableTabs":false,"tabWidth":65,"tabXOffset":10,"hasSecondTab":false,"secondTabText":"Send Back","secondTabOffset":80,"secondTabWidth":65},"widgets_values":[]},{"id":57,"type":"CheckpointLoaderSimple","pos":[119.18760941617633,322.3139964245275],"size":[270,98],"flags":{},"order":0,"mode":0,"inputs":[],"outputs":[{"name":"MODEL","type":"MODEL","links":[66]},{"name":"CLIP","type":"CLIP","links":[64]},{"name":"VAE","type":"VAE","links":[63]}],"properties":{"Node name for S&R":"CheckpointLoaderSimple"},"widgets_values":["zImageTurboAnime_aioBF16.safetensors"]},{"id":43,"type":"VAEDecode","pos":[896.7425171608414,284.41917913282174],"size":[300.5818357492219,52.67445832978015],"flags":{},"order":5,"mode":0,"inputs":[{"name":"samples","type":"LATENT","link":38},{"name":"vae","type":"VAE","link":63}],"outputs":[{"name":"IMAGE","type":"IMAGE","slot_index":0,"links":[45]}],"properties":{"Node name for S&R":"VAEDecode","cnr_id":"comfy-core","ver":"0.3.64","enableTabs":false,"tabWidth":65,"tabXOffset":10,"hasSecondTab":false,"secondTabText":"Send Back","secondTabOffset":80,"secondTabWidth":65},"widgets_values":[]},{"id":44,"type":"KSampler","pos":[897.403374808238,380.4033236373366],"size":[300,390],"flags":{},"order":4,"mode":0,"inputs":[{"name":"model","type":"MODEL","link":66},{"name":"positive","type":"CONDITIONING","link":41},{"name":"negative","type":"CONDITIONING","link":42},{"name":"latent_image","type":"LATENT","link":43}],"outputs":[{"name":"LATENT","type":"LATENT","slot_index":0,"links":[38]}],"properties":{"Node name for S&R":"KSampler","cnr_id":"comfy-core","ver":"0.3.64","enableTabs":false,"tabWidth":65,"tabXOffset":10,"hasSecondTab":false,"secondTabText":"Send Back","secondTabOffset":80,"secondTabWidth":65},"widgets_values":[686212419473868,"randomize",9,1,"euler_ancestral","beta",1]},{"id":9,"type":"SaveImage","pos":[1219.0231829247928,287.6512935914231],"size":[361.3498893344904,477.71118609245843],"flags":{},"order":6,"mode":0,"inputs":[{"name":"images","type":"IMAGE","link":45}],"outputs":[],"properties":{"Node name for S&R":"SaveImage","cnr_id":"comfy-core","ver":"0.3.64","enableTabs":false,"tabWidth":65,"tabXOffset":10,"hasSecondTab":false,"secondTabText":"Send Back","secondTabOffset":80,"secondTabWidth":65},"widgets_values":["Z-Image-Turbo-Anime/ComfyUI"]},{"id":41,"type":"EmptySD3LatentImage","pos":[110,514.7991422743903],"size":[291.1505813587138,106],"flags":{"collapsed":false},"order":1,"mode":0,"inputs":[],"outputs":[{"name":"LATENT","type":"LATENT","slot_index":0,"links":[43]}],"properties":{"Node name for S&R":"EmptySD3LatentImage","cnr_id":"comfy-core","ver":"0.3.64","enableTabs":false,"tabWidth":65,"tabXOffset":10,"hasSecondTab":false,"secondTabText":"Send Back","secondTabOffset":80,"secondTabWidth":65},"widgets_values":[800,1200,2]},{"id":45,"type":"CLIPTextEncode","pos":[447.9563923143061,312.89033991931956],"size":[410,370],"flags":{},"order":2,"mode":0,"inputs":[{"name":"clip","type":"CLIP","link":64}],"outputs":[{"name":"CONDITIONING","type":"CONDITIONING","links":[36,41]}],"properties":{"Node name for S&R":"CLIPTextEncode","cnr_id":"comfy-core","ver":"0.3.73","enableTabs":false,"tabWidth":65,"tabXOffset":10,"hasSecondTab":false,"secondTabText":"Send Back","secondTabOffset":80,"secondTabWidth":65},"widgets_values":["anime style illustrated full body portrait of a teenage girl with long flowing silver hair and soft blue eyes, wearing a light summer dress, gently holding the hem of her skirt as she smiles shyly, standing in a sunlit meadow filled with wildflowers, delicate linework and painterly background, calm and airy atmosphere, detailed illustration style"],"color":"#232","bgcolor":"#353"}],"links":[[36,45,0,42,0,"CONDITIONING"],[38,44,0,43,0,"LATENT"],[41,45,0,44,1,"CONDITIONING"],[42,42,0,44,2,"CONDITIONING"],[43,41,0,44,3,"LATENT"],[45,43,0,9,0,"IMAGE"],[63,57,2,43,1,"VAE"],[64,57,1,45,0,"CLIP"],[66,57,0,44,0,"MODEL"]],"groups":[{"id":2,"title":"Step2 - 影像尺寸與批次數量","bounding":[100,444.79914227439036,310,190],"color":"#3f789e","font_size":24,"flags":{}},{"id":3,"title":"Step3 - 提示詞","bounding":[427.6925821298544,242.30743987573302,450,540],"color":"#3f789e","font_size":24,"flags":{}},{"id":4,"title":"Step1 - 載入模型","bounding":[100.3865960143639,242.30746188132042,310,190],"color":"#3f789e","font_size":24,"flags":{}}],"config":{},"extra":{"ds":{"scale":0.922922342084661,"offset":[10.512598479453185,-21.49914796721918]},"frontendVersion":"1.34.9","workflowRendererVersion":"LG","VHS_latentpreview":false,"VHS_latentpreviewrate":0,"VHS_MetadataImage":true,"VHS_KeepIntermediate":true},"version":0.4}
匯入工作流後若提示遺失節點,建議使用 ComfyUI-Manager 搜尋安裝相關套件。安裝完成後務必重啟 ComfyUI,節點才會出現在清單中。

▲ ComfyUI 的工作流截圖。
整體工作流可以拆成四個階段:
- Step 1 – 載入 Z-Image-Turbo-Anime AIO 模型
- Step 2 – 建立輸出尺寸與批次數量
- Step 3 – 撰寫自然語言動漫提示詞
- Step 4 – 取樣、解碼並輸出圖片
雖然節點數量不多,但模型一次完成了語意理解、構圖決策與動漫風格渲染,這也是整條工作流能維持極簡的關鍵。
示範與生成成果
以下示範展示的是使用 Z-Image-Turbo-Anime 在本文提供之工作流與取樣設定下,僅透過「自然語言提示詞差異」所生成的多組角色插圖。可以看到,不論是少女或少年角色,模型都能穩定維持一致的動漫比例、乾淨的線條風格與完整的全身構圖,同時準確回應提示詞中對於姿勢、表情、服裝與場景氣氛的描述。即使在 Batch size 2、9 steps 的設定下,角色輪廓與畫面完成度仍相當一致,適合用於角色概念設計、插畫草圖或快速風格測試等高頻創作情境。實測一次生成兩張圖大概要花 18 至 19 秒,VRAM 大概吃到 15 GB 多。

▲ 示範一,完整提示詞:
anime style illustrated full body portrait of a teenage girl with long flowing silver hair and soft blue eyes, wearing a light summer dress, gently holding the hem of her skirt as she smiles shyly, standing in a sunlit meadow filled with wildflowers, delicate linework and painterly background, calm and airy atmosphere, detailed illustration style

▲ 示範二,完整提示詞:
anime style illustrated full body portrait of a young girl with short wavy chestnut hair and warm amber eyes, wearing a casual hoodie and skirt, stretching her arms playfully with a bright cheerful smile, standing on a quiet residential street in the afternoon, clean background details, lively and friendly mood, hand-painted illustration look

▲ 示範三,完整提示詞:
anime style illustrated full body portrait of a girl with long straight black hair and deep violet eyes, wearing a traditional-inspired modern outfit, standing gracefully with one hand resting near her chest, gentle and mysterious expression, detailed shrine background at dusk, soft lighting and elegant illustration style

▲ 示範四,完整提示詞:
anime style illustrated full body portrait of a teenage girl with pastel pink twin tails and sparkling green eyes, wearing a stylish school uniform, leaning slightly forward with hands behind her back, playful smile and energetic pose, detailed cherry blossom park background, vibrant and cute illustrated atmosphere

▲ 示範五,完整提示詞:
anime style illustrated full body portrait of a girl with shoulder-length blonde hair and clear sky-blue eyes, wearing a long coat and boots, calmly looking to the side with a thoughtful expression, hands in pockets, standing near a quiet lakeside in autumn, refined color palette and painterly illustration background

▲ 示範六,完整提示詞:
anime style illustrated full body portrait of a teenage boy with messy dark brown hair and deep green eyes, wearing a casual jacket and jeans, standing relaxed with one hand in his pocket, gentle smile on his face, detailed urban street background, clean and modern illustrated look

▲ 示範七,完整提示詞:
anime style illustrated full body portrait of a boy with short silver-gray hair and sharp blue eyes, wearing a simple shirt and trousers, arms crossed confidently, calm and composed expression, standing in front of a minimalist interior with soft light, refined illustration style

▲ 示範八,完整提示詞:
anime style illustrated full body portrait of a young boy with medium-length navy blue hair and golden eyes, wearing a school uniform, adjusting his tie with a slightly embarrassed smile, detailed classroom background with warm afternoon light, expressive character-focused illustration

▲ 示範九,完整提示詞:
anime style illustrated full body portrait of a teenage boy with tousled auburn hair and brown eyes, wearing a scarf and long coat, walking forward with a determined expression, coat flowing in the wind, detailed town street background in early evening, cinematic illustrated atmosphere

▲ 示範十,完整提示詞:
anime style illustrated full body portrait of a boy with straight black hair and soft gray eyes, wearing comfortable everyday clothing, holding a book close to his chest, gentle and introspective expression, standing in a quiet library with detailed shelves behind him, calm and elegant illustration style
結語
Z-Image-Turbo-Anime 把動漫插圖生成這件事,從「模型選擇與參數調校」簡化成一條乾淨、可重複的標準流程。透過 Anime LoRA 的深度合併與 AIO 封裝,它在 ComfyUI 中更像是一個專為二次元創作打造的高速引擎:好載入、好管理,也更容易長期使用。
如果你追求的是單模型、低步數、流程穩定的動漫插圖生成體驗,那麼這條 ComfyUI × Z-Image-Turbo-Anime 工作流,會是一個非常適合收進常用範本的選擇。
《上一篇》AI 繪圖技術怎麼演進?U-Net 架構到 Transformer 時代的關鍵轉變
《下一篇》Z-Image-Turbo-Anime:測試生成高畫質二次元動漫壁紙 









留言區 / Comments
萌芽論壇