encountering a RuntimeError due to NaN values being detected in the latents during training

Traceback (most recent call last):
File "C:\AI\ComfyUI\lora-scripts-v1.12.0\scripts\stable\sdxl_train_network.py", line 185, in
trainer.train(args)
File "C:\AI\ComfyUI\lora-scripts-v1.12.0\scripts\stable\train_network.py", line 272, in train
train_dataset_group.cache_latents(vae, args.vae_batch_size, args.cache_latents_to_disk, accelerator.is_main_process)
File "C:\AI\ComfyUI\lora-scripts-v1.12.0\scripts\stable\library\train_util.py", line 2196, in cache_latents
dataset.cache_latents(vae, vae_batch_size, cache_to_disk, is_main_process)
File "C:\AI\ComfyUI\lora-scripts-v1.12.0\scripts\stable\library\train_util.py", line 1093, in cache_latents
cache_batch_latents(vae, cache_to_disk, batch, condition.flip_aug, condition.alpha_mask, condition.random_crop)
File "C:\AI\ComfyUI\lora-scripts-v1.12.0\scripts\stable\library\train_util.py", line 2592, in cache_batch_latents
raise RuntimeError(f"NaN detected in latents: {info.absolute_path}")
RuntimeError: NaN detected in latents: C:\aa\image\40_aa\1.png
09:48:41-274594 ERROR Training failed / 训练失败
i need help