PixAI会員

📢 ＜ログイン障害＞お詫びと復旧のお知らせ — このたびは、一部のユーザーの皆さまがPixAIアカウントから予期せずログアウトされるという問題が発生し、ご不便をおかけしましたことを、心よりお詫び申し上げます。現在は復旧しております。 ログインできない場合は、お手数ですが再度ログインをお試…

詳細を見る

ユーザーLoRA

XL

Knife XL FFusion - CivitaI / LoRA + FA Text Encoder

#ナイフ
Knife XL FFusion - CivitaI / LoRA + FA Text Encoder - AI Modelのカバー画像
アップロード日時
2025/06/11 3:39
採用数
447
レビュー
レビューがありません
トリガーワード

knife

権限
  • 画像生成と共有を許可する
  • ユーザーによるモデルのダウンロードを許可する
  • 生成画像の商用利用を許可する

おすすめLoRA

ロル - AI Modelのカバー画像

ユーザーLoRA

ロル
5
Null bulge - AI Modelのカバー画像

ユーザーLoRA

Null bulge
0
Aki Murasame (むらさめ アキ) - Mecha-Ude: Mechanical Arms (メカウデ) - AI Modelのカバー画像

ユーザーLoRA

Aki Murasame (むらさめ アキ) - Mecha-Ude: Mechanical Arms (メカウデ)
6
衣装LoRA タコ着ぐるみ衣装 PixAI版LoRA - AI Modelのカバー画像

ユーザーLoRA

衣装LoRA タコ着ぐるみ衣装 PixAI版LoRA
10

詳細説明

🗡️ FFusionAI's Knife LoRA Model DemonstrationsThis demonstration dives deep into the intricacies of three distinct LoRA trainings. Each model has been meticulously trained on a unique dataset of 200 knives, captured in the professional environment of our partner, NoramePhotography Studio. The visuals span from pure white studio shots to rustic wood settings, glinting coins, and fantasy-inspired indoor decor.🔍 Dataset Insight: The dataset, although rich in variety, was curated with a fast and informal tagging approach, mainly for demonstration purposes. If you're intrigued by the knife photo session and wish for a more in-depth training, do let us know!While depth variations are in the pipeline, our current focus revolves around evaluating the distinct LoRA variations.🎯 Models at a Glance:1. CivitAI's Quick LoRA Training (Lora1)📌 Highlights:Powered by CivitAI's new LoRA trainer.Swift 10-epoch run, completed in a breezy 20-30 minutes.Quality may vary with default settings, but hey, time is essence!📊 Specifications:Date: 2023-09-19T14:36:14Resolution: 1024x1024Architecture: stable-diffusion-xl-v1-base/loraNetwork Dim/Rank: 32.0Alpha: 16.0Knife_XL_FFusion.safetensors Date: 2023-09-19T14:36:14 Title: Knife_XL_FFusion Resolution: 1024x1024 Architecture: stable-diffusion-xl-v1-base/lora Network Dim/Rank: 32.0 Alpha: 16.0 Module: networks.lora Learning Rate (LR): 0.0005 UNet LR: 0.0005 TE LR: 5e-05 Optimizer: bitsandbytes.optim.adamw.AdamW8bit(weight_decay=0.1) Scheduler: cosine_with_restarts Warmup steps: 0 Epoch: 10 Batches per epoch: 74 Gradient accumulation steps: 1 Train images: 282 Regularization images: 0 Multires noise iterations: 6.0 Multires noise discount: 0.3 Min SNR gamma: 5.0 Zero terminal SNR: True Max grad norm: 1.0 Clip skip: 1 Dataset dirs: 1 [img] 282 images UNet weight average magnitude: 2.634092236933176 UNet weight average strength: 0.009947009810559605 Text Encoder (1) weight average magnitude: 1.696394163771355 Text Encoder (1) weight average strength: 0.008538951936953606 Text Encoder (2) weight average magnitude: 1.720911101275907 Text Encoder (2) weight average strength: 0.0066990979319423882. LoRA FA with Text Encoder Only (Lora2)📌 Highlights:Exclusive training on text encoder.Absence of UNet in this LoRA variant.📊 Specifications:Date: 2023-09-19T20:04:24Resolution: 1024x1024Architecture: stable-diffusion-xl-v1-base/loraNetwork Dim/Rank: 32.0Alpha: 32.0Knife-FFusion-LoRA-FA.safetensors

Date: 2023-09-19T20:04:24 Title: Knife-FFusion-LoRA-FA Resolution: 1024x1024 Architecture: stable-diffusion-xl-v1-base/lora Network Dim/Rank: 32.0 Alpha: 32.0 Module: networks.lora_fa

Text Encoder (1) weight average magnitude: 3.986337637923385 Text Encoder (1) weight average strength: 0.018590648076750333 Text Encoder (2) weight average magnitude: 4.043434837883338 Text Encoder (2) weight average strength: 0.014620680042179104 No UNet found in this LoRA3. General LoRA Training📌 Highlights:Comprehensive LoRA training with diverse specifications.Trained on an extensive dataset of 485 knife images.📊 Specifications:Date: 2023-08-26T23:08:56Resolution: 1024x1024Architecture: stable-diffusion-xl-v1-base/loraNetwork Dim/Rank: 32.0Alpha: 16.0FF-Minecraft-XL Resolution: 1024x1024 Architecture: stable-diffusion-xl-v1-base/lora Network Dim/Rank: 32.0 Alpha: 16.0 Module: networks.lora Learning Rate (LR): 0.0005 UNet LR: 0.0005 TE LR: 5e-05 Optimizer: bitsandbytes.optim.adamw.AdamW8bit(weight_decay=0.1) Scheduler: cosine_with_restarts Warmup steps: 0 Epoch: 10 Batches per epoch: 121 Gradient accumulation steps: 1 Train images: 458 Regularization images: 0 Multires noise iterations: 6.0 Multires noise discount: 0.3 Min SNR gamma: 5.0 Zero terminal SNR: True Max grad norm: 1.0 Clip skip: 1 Dataset dirs: 1 [img] 458 images UNet weight average magnitude: 2.9987627096874507 UNet weight average strength: 0.011098071585284945 Text Encoder (1) weight average magnitude: 1.729993708156961 Text Encoder (1) weight average strength: 0.008685239007756952 Text Encoder (2) weight average magnitude: 1.7630326984758309 Text Encoder (2) weight average strength: 0.0068346636309082635🎨 Readme Crafted by: 🤖 & FFusionAI 🚀🌐 Contact InformationThe FFusion.ai project is proudly maintained by Source Code Bulgaria Ltd & Black Swan Technologies.📧 Reach us at di@ffusion.ai for any inquiries or support.🌌 Find us on:🐙 GitHub😊 Hugging Face💡 Civitai🌍 Sofia Istanbul London

コメント

Knife XL FFusion - CivitaI / LoRA + FA Text Encoderで生成した作品

1
AI artwork generated with Knife XL FFusion - CivitaI / LoRA + FA Text Encoder
Avatar image of Koom Aton
masterpiece, 1 men, redmod, dramatist, misty, red skin, metal body, red body, bloody body, realistic heart, Arbitrary movement, aesthetic movement, attentive gaze of optical eyes, Demonic nature, holding a heart, holding a medieval style, A bleeding human heart in the palm of your hand, an atmosphere of horror, stiletto Punching through, heart, pubis, network mask on face, (((bottom:1.3))), DLSS 4, Multi Frame Generation, Ray Traicing, Path Tracing, armor elements are missing, (((Almost transparent body:1, 2))), (((mechanical parts inside:1, 3))), knife, <lora:more_details:0.5>, Aeolian, beauty, dark girl, hidden in the shadows, a ray of light illuminating her gaze, a gypsy dressed in black, a transparent black cloth mask, assassin girl, huge weapon, wlop 1.5
Koom Aton
AI artwork generated with Knife XL FFusion - CivitaI / LoRA + FA Text Encoder
Avatar image of Koom Aton
masterpiece, 1 men, redmod, dramatist, misty, red skin, metal body, red body, bloody body, realistic heart, Arbitrary movement, aesthetic movement, attentive gaze of optical eyes, Demonic nature, holding a heart, holding a medieval style, A bleeding human heart in the palm of your hand, an atmosphere of horror, stiletto Punching through, heart, pubis, network mask on face, (((bottom:1.3))), DLSS 4, Multi Frame Generation, Ray Traicing, Path Tracing, armor elements are missing, (((Almost transparent body:1, 2))), (((mechanical parts inside:1, 3))), <lora:more_details:0.5>, Aeolian, beauty, dark girl, hidden in the shadows, a ray of light illuminating her gaze, a gypsy dressed in black, a transparent black cloth mask, assassin girl, huge weapon, wlop 1.5, (((knife:1,5)))
Koom Aton
ホームモデル生成マイページ