<h1 id="heading-6">CivitAI Style Fusion🏆LoRAs</h1><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/70a8fd48-f04d-4fdd-a75f-4cba89031c41/width=525/70a8fd48-f04d-4fdd-a75f-4cba89031c41.jpeg" /><h1 id="heading-6"><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6b9e5fb0-0f44-4d32-8675-2b04b555762d/width=525/6b9e5fb0-0f44-4d32-8675-2b04b555762d.jpeg" /></h1>Last update: 🚀 CivitAI Lora5 32DIM Notebook with dataset <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/710152cc-d028-456c-ac17-db4a102a243a/width=525/710152cc-d028-456c-ac17-db4a102a243a.jpeg" />Last update: 🚀 CivitAI Lora3 Configuration - Trained with CivitAI Trainer🚀 Date: 2023-11-10 | Title: CivitAI_64_ALL🔍 Key Specifications:<ul><li>Resolution: 1024x1024</li><li>Architecture: stable-diffusion-xl-v1-base/lora</li><li>Network Dim/Rank: 64.0, Alpha: 1.0</li><li>Module: networks.lora</li><li>Learning Rates: UNet LR &amp; TE LR set to optimal levels</li><li>Optimizer: Advanced AdamW8bit</li><li>Epochs &amp; Training: Intensive 10 epochs with 576 batches</li></ul>📊 Model Stats:<ul><li>UNet Weight: Mag - 7.602, Str - 0.0187</li></ul><pre><code>Resolution: 1024x1024 Architecture: stable-diffusion-xl-v1-base/lora
Network Dim/Rank: 64.0 Alpha: 1.0
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: constant Warmup steps: 0
Epoch: 10 Batches per epoch: 576 Gradient accumulation steps: 1
Train images: 2304 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] 576 images
UNet weight average magnitude: 7.602270778898858
UNet weight average strength: 0.018722912685324843
Text Encoder (1) weight average magnitude: 2.7649271326702607
Text Encoder (1) weight average strength: 0.009535635958680934
Text Encoder (2) weight average magnitude: 2.6905091182810352
Text Encoder (2) weight average strength: 0.007233532415344915</code></pre><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/ffa14b04-02d8-482b-9172-92962a5fabba/width=525/ffa14b04-02d8-482b-9172-92962a5fabba.jpeg" />Delve into FFusionAI's approach to AI-driven style synthesis with our newly released LoRA models. Each model has been developed using CivitAI's official trainer, ensuring precision and quality.<h3 id="heading-135">🛠️ LoRA Model Overview:</h3><ul><li>LoRA 1 - Lite Version: Designed for quick testing, this model utilizes a small dataset for swift style generation, operating with a 32-dimension capacity.</li><li>LoRA 2 - Community Fusion: A robust model developed from over 500+ images, submitted by various users for the CivitAI contest. This iteration also features a 32-dimension capacity.</li><li>LoRA 3 - Enhanced Fidelity: Building upon LoRA 2, this model is further trained with higher dimensions, focusing on improving the overall image quality.</li><li>LoRA 4 - Comprehensive Style Mash: Our expansive dataset of 1400 images represents a confluence of all FFusionAI submissions. This model undergoes additional UNet training to refine and diversify the generated styles.</li></ul><h3 id="heading-81">1. FFusionAI Style Capture &amp; Fusion Showdown LoRA</h3><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/e8e7e0cc-5e5a-4284-8743-7c5f16401f9c/width=525/e8e7e0cc-5e5a-4284-8743-7c5f16401f9c.jpeg" /><h3 id="heading-136">🎨 Dataset and Training:</h3>Included within the package are curated collections accessible at <a target="_new" rel="ugc" href="https://civarchive.com/collections/96636">CivitAI Collections</a>. The training prompts have been crafted with BLIP-2, FLAN-T5-XL, and ViT-H-14.Please note, original prompts were not utilized for training. Instead, intentional modifications were made using blip2-flan-t5-xl &amp; ViT-H-14/laion2b_s32b_b79k to adjust and enhance the training dataset, which can be reviewed <a target="_new" rel="ugc" href="https://civarchive.com/articles/2755">here</a>.<h3 id="heading-137">🔄 Further Information:</h3>For a detailed examination of the training datasets, parameters, and model specifications, professionals and enthusiasts are encouraged to explore the metadata provided within the collection.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6a13a9ce-d408-4a3f-b0cc-6000942f32bc/width=525/6a13a9ce-d408-4a3f-b0cc-6000942f32bc.jpeg" /><ul><li>LORA 2 🚀 CivitAI Configuration Overview - 2023-11-10</li></ul>🚀 Trained with the Official CivitAI Trainer📅 Date: 2023-11-10🖼️ Title: CivitAI_ALL🔍 Resolution: 1024x1024🏗️ Architecture: stable-diffusion-xl-v1-base/lora⚙️ Key Settings:<ul><li>Network Dim/Rank: 32.0</li><li>Alpha: 1.0</li><li>Module: networks.lora</li><li>Learning Rates: UNet LR - 0.0005, TE LR - 5e-05</li><li>Optimizer: AdamW8bit (weight_decay=0.1)</li><li>Epochs &amp; Batches: 10 epochs, 167 batches/epoch</li><li>Train Images: 576</li></ul>📊 Model Stats:<ul><li>UNet Weight: Mag - 3.755, Str - 0.0135</li><li>Text Encoder (1): Mag - 1.833, Str - 0.0091</li><li>Text Encoder (2): Mag - 1.836, Str - 0.0071</li></ul>🏷️ Prominent Tags:<ul><li>Fusion styles, Artgerm, Beeple</li><li>Dark fantasy, Official artwork, Pinup art</li><li>Digital illustration, Fantasy &amp; Sci-fi</li><li>...and over 4500 more!</li></ul><h3 id="heading-47">🌐 <a target="_blank" rel="ugc" href="http://FFusion.ai">FFusion.ai</a> Contact Information</h3>Proudly maintained by Source Code Bulgaria Ltd &amp; Black Swan Technologies.<ul><li>📧 For collaborations, inquiries, or support: <a target="_blank" rel="ugc" href="mailto:di@ffusion.ai">di@ffusion.ai</a></li><li>🌍 Locations: Sofia | Istanbul | London</li></ul>Connect with Us:<ul><li>🐙 <a target="_blank" rel="ugc" href="https://github.com/1e-2">GitHub</a></li><li>😊 <a target="_blank" rel="ugc" href="https://huggingface.co/FFusion">Hugging </a><a target="_new" rel="ugc" href="https://chat.openai.com/c/4440256e-8c09-4f2e-ab25-1c9abfbf67f9#">Face</a></li><li>💡 <a target="_blank" rel="ugc" href="https://civarchive.com/user/idle/models">Civitai</a></li></ul>Our Websites:<ul><li>🌐 <a target="_blank" rel="ugc" href="http://FFusion.ai">FFusion.ai</a></li><li>🌌 <a target="_blank" rel="ugc" href="http://FFAI.eu">FFAI.eu</a></li><li>🔗 <a target="_blank" rel="ugc" href="http://1e-2.com">1e-2.com</a> </li></ul>

CivitAI 🏆 Style Fusion - FFusionAI Entry (+dataset)

Trained on Civitai Trainer only An attempt to expand the book image.<a target="_blank" rel="ugc" href="https://civitai.com/images/3507510￼￼32dim">https://civitai.com/images/3507510 64dim</a> ver + dataset {
 "unetLR": 0.0005,
 "clipSkip": 1,
 "loraType": "lora",
 "keepTokens": 0,
 "networkDim": 64,
 "numRepeats": 8,
 "resolution": 1024,
 "lrScheduler": "cosine_with_restarts",
 "minSnrGamma": 4,
 "targetSteps": 3840,
 "enableBucket": true,
 "networkAlpha": 1,
 "optimizerArgs": "weight_decay=0.1",
 "optimizerType": "AdamW8Bit",
 "textEncoderLR": 0.00005,
 "maxTrainEpochs": 10,
 "shuffleCaption": false,
 "trainBatchSize": 4,
 "flipAugmentation": false,
 "lrSchedulerNumCycles": 3
}

Files with SHA256: 0fd1cf6a1ad33c3846ed34cc2599a7782e2bc9aa14682ef8985be2727886d6fb