ModelsLab/Tracer (Overwatch 2) [FLUX] LoRA - V3.0

Changelog:Steps training increase to 3000.Training based on v1Changelog:Steps training increase to 3000Model training on Linear and Linear Alpha 16 "Network Rank (Dimension)"Training Configuration:Images: 38 (1024x1024 each)Linear: 16Linear Alpha: 16Batch Size: 1Steps: 3000Noise Scheduler: "flowmatch"Optimizer: "adamw8bit"Learning Rate: 1e-4 (0.0001)Save Dtype: float16