Upload README.md with huggingface_hub

README.md

@@ -40,20 +40,23 @@ This is an **image-classification / backbone** checkpoint (`PvtV2ImageClassify`
 
 ```python
 import os
+
 os.environ["KERAS_BACKEND"] = "torch"  # or "jax" / "tensorflow"
 
 from PIL import Image
-import numpy as np
-from zeromodels.models.pvt_v2 import PvtV2ImageClassify, PvtV2Model
+from zeromodels.models.pvt_v2 import PvtV2ImageClassify, PvtV2Model, PvtV2ImageProcessor
 
 model = PvtV2ImageClassify.from_weights("zeromodels/pvt-v2-b3")
-backbone = PvtV2Model.from_weights("zeromodels/pvt-v2-b3", as_backbone=True)
+processor = PvtV2ImageProcessor.from_weights("zeromodels/pvt-v2-b3")
 
-image = Image.open("your_image.jpg").convert("RGB").resize((224, 224))
-x = np.asarray(image, dtype="float32")[None]  # (1, H, W, 3), raw [0, 255]
-print(model(x).shape)  # (1, num_classes)
-feats = backbone(x)
-print(len(feats), [tuple(f.shape) for f in feats])  # 4-stage feature pyramid
+image = Image.open("your_image.jpg").convert("RGB")
+pixels = processor(image)  # resize + normalize (normalization lives in the processor)
+logits = model(pixels, training=False)
+print(logits.shape)  # (1, num_classes)
+
+# Feature extraction: the backbone without the classifier head
+backbone = PvtV2Model.from_weights("zeromodels/pvt-v2-b3", as_backbone=True)
+features = backbone(pixels, training=False)
 ```
 
 Normalization is baked into the graph, so pass raw `[0, 255]` pixels. Load any PVTv2