Upload models/dinov3_hf_extractor.py with huggingface_hub

models/dinov3_hf_extractor.py

@@ -1,5 +1,6 @@
 """DINOv3 feature extraction via HuggingFace, handling CLS and register tokens."""
 
+import os
 import torch
 import torch.nn as nn
 from transformers import AutoImageProcessor, AutoModel
@@ -22,14 +23,15 @@ class DINOv3HFExtractor(nn.Module):
     def __init__(self, repo_id="facebook/dinov3-vitb16-pretrain-lvd1689m", take_last=None, take_indices=None, trainable=False):
         super().__init__()
 
-        self.proc = AutoImageProcessor.from_pretrained(repo_id)
+        token = os.environ.get("HF_TOKEN")
+        self.proc = AutoImageProcessor.from_pretrained(repo_id, token=token)
         
         # Disable resizing/cropping so native 512x512 maps to 32x32 patches
         for k in ("do_resize", "do_center_crop"):
             if hasattr(self.proc, k):
                 setattr(self.proc, k, False)
         
-        self.model = AutoModel.from_pretrained(repo_id)
+        self.model = AutoModel.from_pretrained(repo_id, token=token)
         self.model.config.output_hidden_states = True
         
         # trainable=True is used by the `dino_only` ablation (fine-tune the backbone);