Outputs
Description
An adopted workflow for locally training flux loras on 16gb of vram.
Adopted from here: https://openart.ai/workflows/kaka/flux-lora-training-in-comfyui/mhY7UndLNPLEGNGiy7kJ
Used minimal training data, screen shot of the training data included.
Built-in nodes
Custom nodes
OptimizerConfig
FluxTrainSave
FluxTrainValidate
SomethingToString
AddLabel
VisualizeLoss
PreviewImage
SaveImage
ImageBatchMulti
FluxTrainEnd
ImageConcatFromBatch
GetImageSizeAndCount
LayerUtility: TextJoin
LoadImageListFromDir //Inspire
TrainDatasetGeneralConfig
FluxTrainValidationSettings
WD14Tagger|pysssss
OllamaVision
LayerUtility: String
FluxTrainModelSelect
LayerUtility: ImageTaggerSave
TrainDatasetAdd
InitFluxLoRATraining
FluxTrainLoop
UploadToHuggingFace