Repo check

CITATION.cff

@@ -0,0 +1,54 @@
+cff-version: 1.2.0
+message: "If you use this software, please cite the paper below and this software release."
+title: "Pathology-Foundation-Model-Benchmark"
+version: "0.0.1"
+authors:
+  - family-names: Lüscher
+    given-names: Jérôme
+  - family-names: Koreuber
+    given-names: Nora
+  - family-names: Franzen
+    given-names: Jannik
+  - family-names: Reith
+    given-names: Fabian H.
+  - family-names: Winklmayr
+    given-names: Claudia
+  - family-names: Baumann
+    given-names: Elias
+  - family-names: Schürch
+    given-names: Christian M.
+  - family-names: Kainmüller
+    given-names: Dagmar
+  - family-names: Rumberger
+    given-names: Josef Lorenz
+
+preferred-citation:
+  type: inproceedings
+  title: "Pathocellbench: A comprehensive benchmark for cell phenotyping."
+  authors:
+    - family-names: Lüscher
+      given-names: Jérôme
+    - family-names: Koreuber
+      given-names: Nora
+    - family-names: Franzen
+      given-names: Jannik
+    - family-names: Reith
+      given-names: Fabian H.
+    - family-names: Winklmayr
+      given-names: Claudia
+    - family-names: Baumann
+      given-names: Elias
+    - family-names: Schürch
+      given-names: Christian M.
+    - family-names: Kainmüller
+      given-names: Dagmar
+    - family-names: Rumberger
+      given-names: Josef Lorenz
+  booktitle: "International Conference on Medical Image Computing and Computer-Assisted Intervention"
+  pages: "411-420"
+  publisher: "Springer Nature Switzerland"
+  address: "Cham"
+  year: 2025
+
+repository-code: "https://github.com/Kainmueller-Lab/Pathology-Foundation-Model-Benchmark"
+url: "https://github.com/Kainmueller-Lab/Pathology-Foundation-Model-Benchmark"

README.md

@@ -1,17 +1,32 @@
 # Pathology-Foundation-Model-Benchmark
+
+
 Benchmarking the ability of pathology foundation models to do cell type classification
 
+This repository corresponds to the benchmarking code of the MICCAI 2025 paper:
+
+Lüscher, J., Koreuber, N., Franzen, J., Reith, F.H., Winklmayr, C., Baumann, E., Schürch, C.M., Kainmüller, D. and Rumberger, J.L., 2025, September. Pathocellbench: A comprehensive benchmark for cell phenotyping. In _International Conference on Medical Image Computing and Computer-Assisted Intervention_ (pp. 411-420). Cham: Springer Nature Switzerland.; https://link.springer.com/chapter/10.1007/978-3-032-04981-0_39
+
 Assumes a .env file with the following variables:
 - `HF_TOKEN`: Huggingface API token to download models with access restrictions
 
 ## Installation
 install the environment with:
 `conda env create -f env.yaml`
 
+## GPU requirements
+
+This project is designed to run training and large-model evaluation on NVIDIA GPUs and HPC clusters. Minimal requirements and recommendations:
+
+- **CUDA / PyTorch**: Built and tested with CUDA 11.8-compatible PyTorch (see `pyproject.toml` supplemental wheel for `cu118`). Install a PyTorch build that matches your system's CUDA runtime (CUDA 11.8 recommended).
+- **NVIDIA drivers & libraries**: NVIDIA driver compatible with CUDA 11.8, cuDNN and NCCL installed on the system (or available via container image).
+
 ## Data 
 The data can be downloaded via huggingface:
 https://huggingface.co/datasets/Kainmueller-Lab/phenobench
 
+The dataloaders can be found in the following repository: https://github.com/Kainmueller-Lab/Bio-Image-Datasets
+
 ## Running the Benchmark
 Once you have downloaded the dataset, and installed the required packages, the code can be run with the following command: