Skipping sample image generation if args.sample_prompts file is missing

[blackmagic24]

It would be great to have the ability to skip the generation of sample images during the training process in flux_train_network.py when the file specified by args.sample_prompts does not exist.
This feature would allow users to have more interactive control over when sample images are generated.

Current Behavior (flux_train_network.py):
At the moment, fine-tuning fails if the args.sample_prompts file is renamed, removed, or modified during training.

  File "/github/sd-scripts/flux_train_network.py", line 285, in sample_images
    flux_train_utils.sample_images(
  File "/github/sd-scripts/library/flux_train_utils.py", line 75, in sample_images
    prompts = train_util.load_prompts(args.sample_prompts)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/github/sd-scripts/library/train_util.py", line 6071, in load_prompts
    with open(prompt_file, "r", encoding="utf-8") as f:
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: 'prompt.txt'

Perhaps like in the 'sample_images_common' function (train_util.py)?

logger.info(f"generating sample images at step / サンプル画像生成 ステップ: {steps}")
    if not os.path.isfile(args.sample_prompts):
        logger.error(f"No prompt file / プロンプトファイルがありません: {args.sample_prompts}")
        return

[kohya-ss]

This makes sense. I will update sooner.