Remove deprecated refs (#50374)

remove deprecated refs

Author: zucchini-nlp

src/transformers/models/depth_pro/convert_depth_pro_weights_to_hf.py

@@ -23,7 +23,7 @@
 from transformers import (
     DepthProConfig,
     DepthProForDepthEstimation,
-    DepthProImageProcessorFast,
+    DepthProImageProcessor,
 )
 
 
@@ -201,7 +201,7 @@ def write_model(
 
 
 def write_image_processor(output_dir: str):
-    image_processor = DepthProImageProcessorFast()
+    image_processor = DepthProImageProcessor()
     image_processor.save_pretrained(output_dir)
     return image_processor
 

src/transformers/models/dinov3_convnext/convert_dinov3_convnext_to_hf.py

@@ -27,7 +27,7 @@
 from PIL import Image
 from torchvision import transforms
 
-from transformers import DINOv3ConvNextConfig, DINOv3ConvNextModel, DINOv3ViTImageProcessorFast
+from transformers import DINOv3ConvNextConfig, DINOv3ConvNextModel, DINOv3ViTImageProcessor
 
 
 HUB_MODELS = {
@@ -106,7 +106,7 @@ def get_transform(resize_size: int = 224):
 
 
 def get_image_processor(resize_size: int = 224):
-    return DINOv3ViTImageProcessorFast(
+    return DINOv3ViTImageProcessor(
         do_resize=True,
         size={"height": resize_size, "width": resize_size},
         resample=2,  # BILINEAR

src/transformers/models/dinov3_vit/convert_dinov3_vit_to_hf.py

@@ -27,7 +27,7 @@
 from PIL import Image
 from torchvision import transforms
 
-from transformers import DINOv3ViTConfig, DINOv3ViTImageProcessorFast, DINOv3ViTModel
+from transformers import DINOv3ViTConfig, DINOv3ViTImageProcessor, DINOv3ViTModel
 
 
 HUB_MODELS = {
@@ -220,7 +220,7 @@ def get_transform(resize_size: int = 224):
 
 
 def get_image_processor(resize_size: int = 224):
-    return DINOv3ViTImageProcessorFast(
+    return DINOv3ViTImageProcessor(
         do_resize=True,
         size={"height": resize_size, "width": resize_size},
         resample=2,  # BILINEAR

src/transformers/models/edgetam/convert_edgetam_to_hf.py

@@ -33,7 +33,7 @@
     EdgeTamModel,
     EdgeTamPromptEncoderConfig,
     EdgeTamVisionConfig,
-    Sam2ImageProcessorFast,
+    Sam2ImageProcessor,
     Sam2Processor,
     TimmWrapperConfig,
 )
@@ -195,7 +195,7 @@ def convert_edgetam_checkpoint(model_name, checkpoint_path, pytorch_dump_folder,
     state_dict = torch.load(checkpoint_path, map_location="cpu")["model"]
     state_dict = replace_keys(state_dict)
 
-    image_processor = Sam2ImageProcessorFast()
+    image_processor = Sam2ImageProcessor()
     processor = Sam2Processor(image_processor=image_processor)
     hf_model = EdgeTamModel(config)
     hf_model.eval()

src/transformers/models/edgetam_video/convert_edgetam_video_to_hf.py

@@ -33,7 +33,7 @@
     EdgeTamVideoModel,
     EdgeTamVideoPromptEncoderConfig,
     EdgeTamVisionConfig,
-    Sam2ImageProcessorFast,
+    Sam2ImageProcessor,
     Sam2VideoProcessor,
     Sam2VideoVideoProcessor,
     TimmWrapperConfig,
@@ -238,7 +238,7 @@ def convert_edgetam_checkpoint(model_name, checkpoint_path, pytorch_dump_folder,
     state_dict = torch.load(checkpoint_path, map_location="cpu")["model"]
     state_dict = replace_keys(state_dict)
 
-    image_processor = Sam2ImageProcessorFast()
+    image_processor = Sam2ImageProcessor()
     video_processor = Sam2VideoVideoProcessor()
     processor = Sam2VideoProcessor(image_processor=image_processor, video_processor=video_processor)
     hf_model = EdgeTamVideoModel(config)

src/transformers/models/eomt/convert_eomt_to_hf.py

@@ -21,7 +21,7 @@
 import torch
 from huggingface_hub import snapshot_download
 
-from transformers import EomtConfig, EomtForUniversalSegmentation, EomtImageProcessorFast
+from transformers import EomtConfig, EomtForUniversalSegmentation, EomtImageProcessor
 
 
 # fmt: off
@@ -240,7 +240,7 @@ def convert_model(
     else:
         MAPPINGS.update(MLP_MAPPINGS["vanilla_mlp"])
 
-    processor = EomtImageProcessorFast(size=size, do_split_image=do_split_image, do_pad=do_pad)
+    processor = EomtImageProcessor(size=size, do_split_image=do_split_image, do_pad=do_pad)
 
     # Save the config and processor
     if output_dir:

src/transformers/models/eomt_dinov3/convert_eomt_dinov3_to_hf.py

@@ -44,7 +44,7 @@
 from PIL import Image
 from safetensors.torch import load_file
 
-from transformers import EomtDinov3Config, EomtDinov3ForUniversalSegmentation, EomtImageProcessorFast
+from transformers import EomtDinov3Config, EomtDinov3ForUniversalSegmentation, EomtImageProcessor
 
 
 CAT_URL = "http://images.cocodataset.org/val2017/000000039769.jpg"
@@ -340,7 +340,7 @@ def convert_model(
 
     model.load_state_dict(merged_state_dict, strict=True, assign=True)
 
-    processor = EomtImageProcessorFast(
+    processor = EomtImageProcessor(
         size={"shortest_edge": image_size, "longest_edge": image_size},
         do_split_image=False,
         do_pad=True,
@@ -373,7 +373,7 @@ def convert_model(
         processor.push_to_hub(repo_id=f"nielsr/{repo_name}")
 
 
-def _prepare_image(processor: EomtImageProcessorFast) -> torch.Tensor:
+def _prepare_image(processor: EomtImageProcessor) -> torch.Tensor:
     image = Image.open(requests.get(CAT_URL, stream=True).raw).convert("RGB")
     inputs = processor(images=image, do_normalize=False, return_tensors="pt")
     return inputs.pixel_values
@@ -573,7 +573,7 @@ def _assert_allclose(reference: Iterable[torch.Tensor], actual: Iterable[torch.T
 def verify_conversion(
     *,
     hf_model: EomtDinov3ForUniversalSegmentation,
-    processor: EomtImageProcessorFast,
+    processor: EomtImageProcessor,
     delta_state_dict: dict[str, torch.Tensor],
     backbone_repo_id: str,
     image_size: int,

src/transformers/models/ernie4_5_vl_moe/convert_ernie4_5_vl_moe_to_hf.py

@@ -25,7 +25,7 @@
 from transformers import (
     AutoTokenizer,
     Ernie4_5_VLMoeConfig,
-    Ernie4_5_VLMoeImageProcessorFast,
+    Ernie4_5_VLMoeImageProcessor,
     Ernie4_5_VLMoeProcessor,
     Ernie4_5_VLMoeVideoProcessor,
     LlamaTokenizer,
@@ -314,7 +314,7 @@ def convert_processor(model_path, save_dir):
     copyfile(hf_hub_download(FONT_REPO, FONT_NAME), Path(save_dir, FONT_NAME))
 
     processor = Ernie4_5_VLMoeProcessor(
-        image_processor=Ernie4_5_VLMoeImageProcessorFast(),
+        image_processor=Ernie4_5_VLMoeImageProcessor(),
         tokenizer=tokenizer,
         video_processor=Ernie4_5_VLMoeVideoProcessor(font=str(Path(save_dir, FONT_NAME))),
         chat_template=tokenizer.chat_template,

src/transformers/models/gemma3n/convert_gemma3n_weights.py

@@ -43,7 +43,7 @@
     Gemma3nVisionConfig,
     GemmaTokenizer,
     GenerationConfig,
-    SiglipImageProcessorFast,
+    SiglipImageProcessor,
 )
 from transformers.image_utils import PILImageResampling
 from transformers.tokenization_utils_sentencepiece import SentencePieceExtractor
@@ -761,7 +761,7 @@ def main(*args):
     logging.info("Saved GemmaTokenizer for %s to %s", variant, output_path)
 
     feature_extractor = Gemma3nAudioFeatureExtractor()
-    image_processor = SiglipImageProcessorFast(
+    image_processor = SiglipImageProcessor(
         image_seq_length=256,
         image_mean=(0.5,) * 3,
         image_std=(0.5,) * 3,

src/transformers/models/llama4/convert_llama4_weights_to_hf.py

@@ -13,7 +13,7 @@
     GenerationConfig,
     Llama4Config,
     Llama4ForConditionalGeneration,
-    Llama4ImageProcessorFast,
+    Llama4ImageProcessor,
     Llama4Processor,
     Llama4TextConfig,
     Llama4VisionConfig,
@@ -680,7 +680,7 @@ def write_tokenizer(args):
     )
     tokenizer = converter.converted_tokenizer
 
-    image_processor = Llama4ImageProcessorFast()
+    image_processor = Llama4ImageProcessor()
     processor = Llama4Processor(
         image_processor=image_processor,
         tokenizer=tokenizer,