Upload folder using huggingface_hub

README.md

@@ -0,0 +1,99 @@
+---
+pretty_name: "SIB-CL Datasets"
+license: apache-2.0
+tags:
+- photonic-crystal
+- quantum
+- contrastive-learning
+- scientific-machine-learning
+- hdf5
+- pytorch
+- science
+- physics
+---
+
+# SIB-CL Datasets
+
+This repository contains the **Surrogate- and Invariance-Boosted Contrastive Learning (SIB-CL)** datasets for two scientific problems:
+
+* **PhC2D**: 2D photonic crystal density-of-states (DOS) and bandstructure data.
+* **TISE**: 3D time-independent Schrödinger equation eigenvalue and eigenvector solutions.
+
+The data and loader scripts reproduce the behavior of the original PyTorch `Dataset` classes from the SIB-CL paper:
+
+> Loh, C., Christensen, T., Dangovski, R., Kim, S., & Soljačić, M. (2022). "Surrogate- and invariance-boosted contrastive learning for data-scarce applications in science." *Nature Communications*, 13, 4223. [https://www.nature.com/articles/s41467-022-31915-y](https://www.nature.com/articles/s41467-022-31915-y)
+
+## Repository Structure
+
+```
+sibcl-dataset/
+├── phc2d.h5
+├── tise.h5
+├── dataset_loader.py
+└── README.md
+```
+
+* `phc2d.h5`: HDF5 with groups:
+
+  * `unitcell/mpbepsimage/<index>`: 32×32 permittivity images
+  * `unitcell/epsavg/<index>`: scalar average permittivity
+  * `mpbcal/DOS/<index>`, `mpbcal/DOSeldiff/<index>`: DOS arrays
+  * `mpbcal/efreq/`, `mpbcal/emptylattice/`, `mpbcal/eldiff/` for bandstructure
+
+* `tise.h5`: HDF5 with groups:
+
+  * `unitcell/potential_fil` (or `_res<downres>`): potential arrays
+  * `eigval_fil/<index>`: eigenvalues
+  * `eigvec_fil/<index>`: eigenvectors
+
+* `dataset_loader.py`: Hugging Face `datasets` script implementing both PhC2D and TISE loaders, with full config options.
+
+## Metadata
+
+```yaml
+dataset_name: sibcl-dataset
+builder_name: SIBCLDataset
+configs:
+  - name: phc2d
+    description: "2D photonic crystal DOS & bandstructure"
+  - name: tise
+    description: "3D TISE eigenvalue & eigenvector solutions"
+license: apache-2.0
+features:
+  phc2d:
+    inputeps: Array3D(shape=(1,32,32), dtype=float32)
+    y: Sequence(Value(float32))
+    EL: Sequence(Value(float32))
+    ELdiff: Sequence(Value(float32))
+    epsavg: Value(float32)
+  tise:
+    input: Array3D(shape=(1,<downres>,<downres>), dtype=float32)
+    y: Sequence(Value(float32))
+splits:
+  train: 10000
+  validation: 500
+  test: 2000
+```
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+# Load PhC2D (DOS prediction)
+ds_phc = load_dataset("YOUR_USERNAME/sibcl-dataset", name="phc2d", split="train")
+
+# Load TISE (eigenvalue prediction)
+ds_tise = load_dataset("YOUR_USERNAME/sibcl-dataset", name="tise", split="test")
+```
+
+## Processing & Training
+
+For processing pipelines and model training code, see the SIB-CL GitHub repository:
+[https://github.com/clott3/SIB-CL](https://github.com/clott3/SIB-CL)
+
+---
+
+**Citation**:
+Loh, C., Christensen, T., Dangovski, R., Kim, S., & Soljačić, M. (2022). "Surrogate- and invariance-boosted contrastive learning for data-scarce applications in science." *Nature Communications*, 13, 4223. [https://www.nature.com/articles/s41467-022-31915-y](https://www.nature.com/articles/s41467-022-31915-y)
+

cylin-tm-11k.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18f4a8e82ba901d696c9a763c84f14d62cad300b89968e310fff92cf2c064e39
+size 1863836500

mf1-tm-32k.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6fa0fd6cddd2e7955bab07acd2cf1608febd12bf375894309e5e28e5ddfafca
+size 5770918152

tise_3d_mf2_sigfac20_xmax5_me1.0_e0_n32_LR4_35k.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae10ad122a09cd535645fe4252b9466875bba026da4a9dd1bc48981d63fd8818
+size 41650599460

tise_3d_mf2_sigfac20_xmax5_me1.0_e0_n32_LR5.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f32738083c14d2236118658ebac7323cbf331b3858985791957e9f453f57c745
+size 17857611632

tise_3d_mf2_sigfac20_xmax5_me1.0_e0_n32_res64.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9639d97f8d19a8290ea3f3ca8577d52cc3c513b315aee92ee376f7681d160bb1
+size 37228324112