note: you might need to use roop or Adetailer to fix faces , as the training data is old blurry 8mm video.
retro vintage 8mm SDXL Lora, trained on old 1960s 8mm film stills.
