ControlNetCannyReducedImproved / training_config.yaml
tommycik's picture
Upload training_config.yaml with huggingface_hub
06cf9e7 verified
N4: true
checkpointing_steps: 250
controlnet_model: InstantX/FLUX.1-dev-Controlnet-Canny
controlnet_type: canny
gradient_accumulation_steps: 4
hub_model_id: tommycik/ControlNetCannyReducedImproved
learning_rate: 5e-06
mixed_precision: 'no'
resolution: 512
steps: 2000
train_batch_size: 1
validation_image: None
validation_steps: 125