Upload pretrained SSL4EM encoders

4dstem/dinov2_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "dinov2",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

4dstem/dinov2_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d82ba4de4bac9644d3ea99f4e4b4e7ea5a88ce52d06e39ac2adee0206c437d0
+size 341673235

4dstem/dinov2_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "dinov2",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

4dstem/dinov2_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f3610166c45cb292009b8695e2f48ee1f39370c28d81f5371f4ab55622c4892
+size 341673235

4dstem/ijepa_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "ijepa",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

4dstem/ijepa_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af57567a06e8c117008e89efe2900bab8d8f12601ff246a1e6d7c71818c85a5a
+size 341673235

4dstem/ijepa_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "ijepa",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

4dstem/ijepa_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3dce8753b4ca20aa7a32f0c23d7520cd1aad44f24b61cf545d91780cf99a7f1
+size 341673235

4dstem/mae_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "mae",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

4dstem/mae_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c303bf67a9dcec77824cb866db2fe242c7575860bd599d83768c8abd991d6a1b
+size 341673235

4dstem/mae_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "mae",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

4dstem/mae_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29bbc2e1e773bc8d5a8ff866211f868210b8cc72bb97e09d3d9a60be7c2f99d5
+size 341673235

4dstem/simclr_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "simclr",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

4dstem/simclr_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f78a1033646df7b7e6173bdf1ad1a999b620800bc8186e6626995e7aeb2d4a7
+size 341673235

4dstem/simclr_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "simclr",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

4dstem/simclr_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61937faf15f525fb36da0a0cceca8587c1a240fa7f216245060c70aaee6b295d
+size 341673235

4dstem/vicregl_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "vicregl",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

4dstem/vicregl_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fce8be01480a550526e3adad270665164520df6e79dd0d291164c21c0e0fb83c
+size 341673235

4dstem/vicregl_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 224,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "vicregl",
+  "pretrain_dataset": "4dstem",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

4dstem/vicregl_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7552dcbc2ea08baa41988cdbe3e645afc051a28d6add849b7076d6c0537ef938
+size 341673235

README.md

@@ -0,0 +1,110 @@
+---
+license: mit
+library_name: pytorch
+pipeline_tag: image-feature-extraction
+tags:
+- self-supervised-learning
+- electron-microscopy
+- 4d-stem
+- scientific-imaging
+- dinov2
+- mae
+- simclr
+- vicregl
+- i-jepa
+---
+
+# Physics-Aligned Self-Supervised Learning for Scientific Imaging
+
+Pretrained ViT-B encoders from the GCPR 2026 paper
+*Physics-Aligned Self-Supervised Learning for Scientific Imaging*.
+
+Five SSL methods (DINOv2, I-JEPA, MAE, SimCLR, VICRegL) were each pretrained
+on two electron-microscopy modalities under two augmentation regimes:
+
+- **`*_domain`** — physics-aligned augmentations (T_phys): measurement-consistent
+  symmetries plus acquisition-driven perturbations (noise, intensity variation,
+  reciprocal-space scaling, diffraction tilt, ...).
+- **`*_original`** — standard natural-image augmentations (T_orig): random crop,
+  horizontal flip, blur, photometric perturbations.
+
+Pretraining data:
+
+- **`cem500k/`** — real-space cellular EM ([CEM500K](https://doi.org/10.7554/eLife.65894) subset, 10k images).
+- **`4dstem/`** — simulated LiNiO2 4D-STEM diffraction patterns
+  ([Scheunert et al.](https://doi.org/10.5281/zenodo.17360572) subset, 10k patterns).
+
+All encoders are single-channel (grayscale) ViT-B backbones.
+
+## Available models
+
+| Path | Method | Pretraining data | Augmentations |
+|---|---|---|---|
+| `cem500k/dinov2_domain` | dinov2 | cem500k | physics-aligned |
+| `cem500k/dinov2_original` | dinov2 | cem500k | natural-image |
+| `cem500k/ijepa_domain` | ijepa | cem500k | physics-aligned |
+| `cem500k/ijepa_original` | ijepa | cem500k | natural-image |
+| `cem500k/mae_domain` | MAE | cem500k | physics-aligned |
+| `cem500k/mae_original` | MAE | cem500k | natural-image |
+| `cem500k/simclr_domain` | simclr | cem500k | physics-aligned |
+| `cem500k/simclr_original` | simclr | cem500k | natural-image |
+| `cem500k/vicregl_domain` | vicregl | cem500k | physics-aligned |
+| `cem500k/vicregl_original` | vicregl | cem500k | natural-image |
+| `4dstem/dinov2_domain` | dinov2 | 4dstem | physics-aligned |
+| `4dstem/dinov2_original` | dinov2 | 4dstem | natural-image |
+| `4dstem/ijepa_domain` | ijepa | 4dstem | physics-aligned |
+| `4dstem/ijepa_original` | ijepa | 4dstem | natural-image |
+| `4dstem/mae_domain` | MAE | 4dstem | physics-aligned |
+| `4dstem/mae_original` | MAE | 4dstem | natural-image |
+| `4dstem/simclr_domain` | simclr | 4dstem | physics-aligned |
+| `4dstem/simclr_original` | simclr | 4dstem | natural-image |
+| `4dstem/vicregl_domain` | vicregl | 4dstem | physics-aligned |
+| `4dstem/vicregl_original` | vicregl | 4dstem | natural-image |
+
+## Usage
+
+With the accompanying code (https://github.com/DL4EM/physics-aligned-ssl):
+
+```python
+from em_ssl.hub import load_encoder
+
+encoder = load_encoder("cem500k/dinov2_domain", repo_id="DL4EM/physics-aligned-ssl")
+
+import torch
+images = torch.randn(4, 1, 128, 128)   # grayscale EM crops in [0, 1]
+features = encoder(images)
+```
+
+Without the codebase, each `encoder.pt` is a plain PyTorch checkpoint:
+
+```python
+import torch
+from huggingface_hub import hf_hub_download
+
+path = hf_hub_download("DL4EM/physics-aligned-ssl", "cem500k/dinov2_domain/encoder.pt")
+ckpt = torch.load(path, map_location="cpu", weights_only=True)
+state_dict = ckpt["encoder_state_dict"]     # ViT-B weights
+print(ckpt["backbone"], ckpt["backbone_kwargs"])
+```
+
+Inputs are single-channel images normalised to [0, 1] (percentile
+normalisation was used during pretraining). Real-space EM models were trained
+on 128x128 crops; 4D-STEM models on 224x224 crops.
+
+## Citation
+
+```bibtex
+@inproceedings{kazimi2026physicsaligned,
+  title     = {Physics-Aligned Self-Supervised Learning for Scientific Imaging},
+  author    = {Kazimi, Bashir and Sandfeld, Stefan},
+  booktitle = {DAGM German Conference on Pattern Recognition (GCPR)},
+  year      = {2026}
+}
+```
+
+## About
+
+Developed by the [Deep Learning for Electron Microscopy (DL4EM)](https://www.fz-juelich.de/en/ias/ias-9/research/deep-learning-for-electron-microscopy)
+group at the [Institute for Materials Data Science and Informatics (IAS-9)](https://www.fz-juelich.de/en/ias/ias-9),
+Forschungszentrum Jülich. For questions, please open an issue on the
+[GitHub repository](https://github.com/DL4EM/physics-aligned-ssl).

cem500k/dinov2_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "dinov2",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

cem500k/dinov2_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb63bbfcedf1cde172f4803705908495bbbe5d2efc79bcef43063ea60f759ce7
+size 341267731

cem500k/dinov2_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "dinov2",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

cem500k/dinov2_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b22e9f60f6cf4c0f879737b6a54bdd800c926e6f966e4d4b334319049c9b110
+size 341267731

cem500k/ijepa_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 8,
+    "drop_rate": 0.0
+  },
+  "method": "ijepa",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

cem500k/ijepa_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e381214596bdf96fc0ba8eff4f56194cc55ca75f4283668aa7b8b4d9acbc8d66
+size 341267731

cem500k/ijepa_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 8,
+    "drop_rate": 0.0
+  },
+  "method": "ijepa",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

cem500k/ijepa_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d244356597d545afc4a10a90219ac4fc8b8352cba72641483e497a6847ce862b
+size 341267731

cem500k/mae_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "mae",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

cem500k/mae_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90696236832e3074e4ca7b12fe15e0300c6f7a2fb85bb98fcdcfb6aae4e1b08d
+size 341267731

cem500k/mae_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "mae",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

cem500k/mae_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26142cf7fae04a8d908c7aeae29c1ecd25de5ec8cb0f625e0d56a53d70e22fe6
+size 341267731

cem500k/simclr_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "simclr",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

cem500k/simclr_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f23081d72c85667552e2daa390ff97d1720ff5444c3e488e076ac5756f362db9
+size 341267731

cem500k/simclr_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "simclr",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

cem500k/simclr_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7fce96ffb7a87cf605bc4c1675ba5c40922ff70be54e470238005580982daff3
+size 341267731

cem500k/vicregl_domain/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "vicregl",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "physics-aligned",
+  "epoch": 299
+}

cem500k/vicregl_domain/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f5097420a0d97858d786850369b17ffa0dc054bae282ff1f71287e1e707d4cf
+size 341267731

cem500k/vicregl_original/config.json

@@ -0,0 +1,13 @@
+{
+  "backbone": "vit_base_em",
+  "in_channels": 1,
+  "backbone_kwargs": {
+    "img_size": 128,
+    "patch_size": 16,
+    "drop_rate": 0.0
+  },
+  "method": "vicregl",
+  "pretrain_dataset": "cem500k",
+  "augmentations": "natural-image",
+  "epoch": 299
+}

cem500k/vicregl_original/encoder.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6601446f639cdaff8a773148229076dbe5956278d150771df48eec3406b3828d
+size 341267731