fix: use num_labels=1 instead of num_classes=1

num_classes=1 maps to num_labels=2 internally in ViTConfig,
creating [2,384] classifier that mismatches the safetensors.
num_labels=1 produces [1,384] — correct for single-class sigmoid.

README.md

@@ -38,7 +38,7 @@ The `model.safetensors` has been regenerated from the correct training checkpoin
 | Wrong `model.safetensors` | Weights from different model (intermediate_size=3072, wrong classifier) | Regenerated from `pretrained_weights/model_v11_ViT_384_base_ckpt.pt` |
 | `num_attention_heads: 12` | **Silently wrong** — attention sliced 12×32d instead of 6×64d | `6` |
 | Preprocessor `size` | Squashed non-square images or no center-crop | `shortest_edge: 440` + `do_center_crop` |
-| `num_classes: 2` / missing `id2label` | Wrong output format for single-class classifier | `1` (sigmoid output) |
+| `num_classes: 2` / no `num_labels` | Wrong output format for single-class classifier — `num_classes=1` maps to 2 labels internally | `num_labels: 1` (sigmoid output) |
 
 ### ⚠️ Breaking change for older transformers versions
 
@@ -50,7 +50,7 @@ This model now requires `transformers >= 5.4.0` for correct image preprocessing.
 import json
 with open("path/to/config.json") as f:
     cfg = json.load(f)
-assert cfg["num_classes"] == 1, "Still broken — re-download the model"
+assert cfg["num_labels"] == 1, "Still broken — re-download the model"
 assert cfg["num_attention_heads"] == 6, "Still broken — re-download the model"
 assert cfg["intermediate_size"] == 1536, "Still broken — re-download the model"
 ```

config.json

@@ -14,7 +14,7 @@
   "model_type": "vit",
   "num_attention_heads": 6,
   "num_channels": 3,
-  "num_classes": 1,
+  "num_labels": 1,
   "num_hidden_layers": 12,
   "patch_size": 16,
   "qkv_bias": true,