Upload anytext.py

anytext.py

@@ -58,6 +58,8 @@ from diffusers.pipelines.pipeline_utils import DiffusionPipeline, StableDiffusio
 from diffusers.pipelines.stable_diffusion.pipeline_output import StableDiffusionPipelineOutput
 from diffusers.pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 from diffusers.schedulers import KarrasDiffusionSchedulers
+from diffusers.configuration_utils import register_to_config, ConfigMixin
+from diffusers.models.modeling_utils import ModelMixin
 from diffusers.utils import (
     USE_PEFT_BACKEND,
     deprecate,
@@ -735,7 +737,8 @@ class FrozenCLIPEmbedderT3(AbstractEncoder):
         return self
 
 
-class TextEmbeddingModule(nn.Module):
+class TextEmbeddingModule(ModelMixin, ConfigMixin):
+    @register_to_config
     def __init__(self, font_path, use_fp16=False, device="cpu"):
         super().__init__()
         self.font = ImageFont.truetype(font_path, 60)
@@ -748,7 +751,6 @@ class TextEmbeddingModule(nn.Module):
         args = {}
         args["rec_image_shape"] = "3, 48, 320"
         args["rec_batch_num"] = 6
-        args["rec_char_dict_path"] = "./text_embedding_module/OCR/ppocr_keys_v1.txt"
         args["rec_char_dict_path"] = hf_hub_download(
             repo_id="tolgacangoz/anytext",
             filename="text_embedding_module/OCR/ppocr_keys_v1.txt",