Revert "Allow loading from npz files only without images"

This reverts commit df6712b.

library/train_util.py

@@ -1033,10 +1033,6 @@ def cache_text_encoder_outputs(
             )
 
     def get_image_size(self, image_path):
-        npz_path = os.path.splitext(image_path)[0] + ".npz"
-        if self.is_latent_cacheable and os.path.exists(npz_path):
-            _, size, _, _ = load_latents_from_disk(npz_path)
-            return size
         image = Image.open(image_path)
         return image.size
 
@@ -1456,23 +1452,15 @@ def load_dreambooth_dir(subset: DreamBoothSubset):
                 logger.warning(f"not directory: {subset.image_dir}")
                 return [], []
 
-            img_paths = glob_images(subset.image_dir, "*", self.is_latent_cacheable)
+            img_paths = glob_images(subset.image_dir, "*")
             logger.info(f"found directory {subset.image_dir} contains {len(img_paths)} image files")
 
             # 画像ファイルごとにプロンプトを読み込み、もしあればそちらを使う
             captions = []
             missing_captions = []
-            cached_captions = []
             for img_path in img_paths:
                 cap_for_img = read_caption(img_path, subset.caption_extension)
                 if cap_for_img is None and subset.class_tokens is None:
-                    if self.is_text_encoder_output_cacheable:
-                        cache_file = os.path.splitext(img_path)[0] + TEXT_ENCODER_OUTPUTS_CACHE_SUFFIX
-                        if os.path.exists(cache_file):
-                            captions.append("")
-                            cached_captions.append(cache_file)
-                        continue
-
                     logger.warning(
                         f"neither caption file nor class tokens are found. use empty caption for {img_path} / キャプションファイルもclass tokenも見つかりませんでした。空のキャプションを使用します: {img_path}"
                     )
@@ -1487,26 +1475,19 @@ def load_dreambooth_dir(subset: DreamBoothSubset):
 
             self.set_tag_frequency(os.path.basename(subset.image_dir), captions)  # タグ頻度を記録
 
-            def show_caption_warning(captions_with_warnings, warning_message):
-                if not captions_with_warnings:
-                    return
-
-                number_of_warning_captions = len(captions_with_warnings)
-                number_of_warning_captions_to_show = 5
-                remaining_warning_captions = number_of_warning_captions - number_of_warning_captions_to_show
-
-                logger.warning(warning_message.format(number_of_warning_captions=number_of_warning_captions))
-                for i, warning_caption in enumerate(captions_with_warnings):
-                    if i >= number_of_warning_captions_to_show:
-                        logger.warning(warning_caption + f"... and {remaining_warning_captions} more")
-                        break
-                    logger.warning(warning_caption)
+            if missing_captions:
+                number_of_missing_captions = len(missing_captions)
+                number_of_missing_captions_to_show = 5
+                remaining_missing_captions = number_of_missing_captions - number_of_missing_captions_to_show
 
-            show_caption_warning(missing_captions, 
-                                 "No caption file found for {number_of_warning_captions} images. Training will continue without captions for these images. If class token exists, it will be used. / {number_of_warning_captions}枚の画像にキャプションファイルが見つかりませんでした。これらの画像についてはキャプションなしで学習を続行します。class tokenが存在する場合はそれを使います。")
-
-            show_caption_warning(cached_captions,
-                                 "No caption file found for {number_of_warning_captions} images. Cached TE embeddings are available for this caption, which will be used instead.")
+                logger.warning(
+                    f"No caption file found for {number_of_missing_captions} images. Training will continue without captions for these images. If class token exists, it will be used. / {number_of_missing_captions}枚の画像にキャプションファイルが見つかりませんでした。これらの画像についてはキャプションなしで学習を続行します。class tokenが存在する場合はそれを使います。"
+                )
+                for i, missing_caption in enumerate(missing_captions):
+                    if i >= number_of_missing_captions_to_show:
+                        logger.warning(missing_caption + f"... and {remaining_missing_captions} more")
+                        break
+                    logger.warning(missing_caption)
             return img_paths, captions
 
         logger.info("prepare images.")
@@ -1884,7 +1865,7 @@ def __init__(
 
         extra_imgs = []
         for subset in subsets:
-            conditioning_img_paths = glob_images(subset.conditioning_data_dir, "*", self.is_latent_cacheable)
+            conditioning_img_paths = glob_images(subset.conditioning_data_dir, "*")
             extra_imgs.extend(
                 [cond_img_path for cond_img_path in conditioning_img_paths if cond_img_path not in cond_imgs_with_img]
             )
@@ -2163,23 +2144,15 @@ def debug_dataset(train_dataset, show_input_ids=False):
 
         epoch += 1
 
-def glob_images(directory, base="*", fallback_to_cache=False):
+
+def glob_images(directory, base="*"):
     img_paths = []
     for ext in IMAGE_EXTENSIONS:
         if base == "*":
             img_paths.extend(glob.glob(os.path.join(glob.escape(directory), base + ext)))
         else:
             img_paths.extend(glob.glob(glob.escape(os.path.join(directory, base + ext))))
     img_paths = list(set(img_paths))  # 重複を排除
-
-    if fallback_to_cache and len(img_paths) == 0:
-        print(f"No images found in {directory}. Will look for cached latents instead.")
-        if base == "*":
-            img_paths.extend(glob.glob(os.path.join(glob.escape(directory), base + ".npz")))
-        else:
-            img_paths.extend(glob.glob(glob.escape(os.path.join(directory, base + ".npz"))))
-
-        img_paths = [img_path for img_path in set(img_paths) if not img_path.endswith(TEXT_ENCODER_OUTPUTS_CACHE_SUFFIX)]
     img_paths.sort()
     return img_paths