Upload folder using huggingface_hub

README.md

@@ -160,6 +160,29 @@ bash eval/run_patch_eval.sh uni crc                 # reference FM
 Reference FM weights are pulled from their Hugging Face hubs on first use (UNI / UNI2-h / Virchow2
 are gated — request access on HF beforehand).
 
+### Evaluate your model on AMC-HCC-ST — we run it for you
+
+AMC-HCC-ST is an in-house, **contamination-free** spatial-transcriptomics cohort that we are actively
+**curating and expanding** at Asan Medical Center. Because it is patient-derived, the cohort **cannot
+be publicly redistributed**. Rather than keep it as an internal-only benchmark, **we offer to run the
+evaluation on your behalf** — send us your pathology encoder and we return its AMC-HCC-ST score under
+the exact protocol used above (leave-one-patient-out ridge, top-50 HVG, mean Pearson), directly
+comparable to the reference models.
+
+**What to send**
+- **Weights** — a `teacher_checkpoint.pth` / `state_dict`, or a public Hugging Face / `timm` hub id.
+- **A loader** — a small `build()` returning an `nn.Module` that maps a normalized `(B, 3, 224, 224)`
+  batch to a `(B, d)` tile embedding (CLS or pooled), plus the expected input normalization
+  (ImageNet by default). See `eval/openpath_eva_backbone.py` for the exact interface we use.
+- **Optional** — a one-line model description and license so we can report your result correctly.
+
+Every submission runs through the same single backbone-factory + probing pipeline (`eval/`), so your
+numbers are apples-to-apples with the table above. This keeps the benchmark **leakage-controlled and
+open to the community** even though the underlying data stays private.
+
+**Contact:** open a discussion on the [`taejoon89/openpath`](https://huggingface.co/taejoon89/openpath)
+model repo, or email the authors *(TODO: add contact address)*.
+
 ## Intended use & limitations
 
 **Intended use.** OpenPath is a **frozen feature extractor** for H&E histopathology. It produces a