pimm export (config.json)

README.md

@@ -1,3 +1,28 @@
 ---
-license: mit
+library_name: pimm
+tags:
+  - particle-physics
+  - lartpc
+  - point-cloud
 ---
+
+# PoLAr-MAE — semantic segmentation
+
+[PoLAr-MAE](https://arxiv.org/abs/2502.02558) fine-tuned for 4-class LArTPC semantic segmentation (shower, track, Michel, delta); reproduces the paper mF1 ≈ 0.82.
+
+- **pimm type:** `PoLArMAE-SemSeg` · 4 classes
+
+> Faithful eval needs PoLAr-MAE preprocessing: `LogTransform(min_val=0.13)`, `energy_threshold=0.13`, `remove_low_energy_scatters=True`. Coordinate normalization is in-model.
+
+## Loading
+
+```python
+import pimm
+model = pimm.from_pretrained("hf://deeplearnphysics/polar-mae-semantic")
+```
+
+Architecture + hyper-parameters travel in `config.json`; weights are bitwise-identical to the original checkpoint.
+
+## Provenance
+
+Repackaged from the original [PoLAr-MAE](https://github.com/DeepLearnPhysics/PoLAr-MAE) release checkpoints into the [pimm](https://github.com/youngsm/particle-imaging-models) export format. Inherits the source repo license.

config.json

@@ -0,0 +1,47 @@
+{
+  "model": {
+    "type": "PoLArMAE-SemSeg",
+    "num_classes": 4,
+    "arch": "vit_small",
+    "voxel_size": 5.0,
+    "num_channels": 4,
+    "seg_head_fetch_layers": [
+      3,
+      7,
+      11
+    ],
+    "seg_head_combination_method": "mean",
+    "seg_head_dim": 512,
+    "seg_head_dropout": 0.5,
+    "freeze_encoder": false,
+    "apply_encoder_postnorm": false,
+    "condition_global_features": true,
+    "upsampling_k": 5,
+    "center": [
+      384.0,
+      384.0,
+      384.0
+    ],
+    "scale": 0.0015035163260146505,
+    "transformer_kwargs": {
+      "postnorm": false,
+      "add_pos_at_every_layer": true,
+      "drop_rate": 0.0,
+      "attn_drop_rate": 0.05,
+      "drop_path_rate": 0.25
+    },
+    "criteria": [
+      {
+        "type": "CrossEntropyLoss",
+        "loss_weight": 1.0,
+        "ignore_index": -1
+      },
+      {
+        "type": "LovaszLoss",
+        "mode": "multiclass",
+        "loss_weight": 0.05,
+        "ignore_index": -1
+      }
+    ]
+  }
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2de651fffcf83e22d388a56f315e031836b86195b022d7c869b4c3993fbd90f
+size 91470268