FF7 Remake version. NSFW dataset.Create your own customized characters with Face Swap and Blend options. One of many features available in my <a target="_blank" rel="ugc" href="https://github.com/dgracey01/SDXL-LoRA-Training-Suite-2.0">Lora Training Suite Project</a>. Come check it out along with my other projects. Come visit my HuggingFace <a target="_blank" rel="ugc" href="https://huggingface.co/spaces/Zero-000/lora-library">Repository</a> for more LoRA’s. Buy me a Ko-Fi will ya!?

Tifa Lockhart

Hybrid training. Both tags and captions via WD EVA02-Large and joyCaption.Specifications:Ranks: 96 (conv:32)Rank Saturated: 100%Base: sdxl_base_v1-0Training Parameters — tifa-trainerxl-step00010800.safetensorsSteps: 10800Epochs: 148Repeats: 1Images: 432Eff. Batch: 2 × 3 = 6Learning Rate: 0.0001UNet LR: 0.0001TE LR: 1e-05Optimizer: AdamWScheduler: cosine_with_restartsLoss: l2Seed: 42Noise Offset: 0.1Min-SNR: 5.0

Files with SHA256: 89bfcaf77f2d83f765556e84a1ea7c637709fdc6dbad75e6b311b957282cc110