Initial upload via tools/push_to_hf.py (architecture: ilex.models.krakencoder.Krakencoder)

README.md

@@ -0,0 +1,67 @@
+---
+library_name: ilex
+tags:
+- jax
+- equinox
+- ilex
+- neuroimaging
+- brain-connectome
+license: mit
+license_link: https://opensource.org/licenses/MIT
+---
+
+# Krakencoder -- joint brain connectome translation autoencoder -- Targlatent singleton -- SCifod2act_coco439_sift2volnorm
+
+## Description
+
+Krakencoder (Jamison, Gu, Wang, Tozlu, Sabuncu, Kuceyeski, *Nature Methods* 2025), ported to JAX / Equinox from the upstream PyTorch release (github.com/kjamison/krakencoder). A linked autoencoder that simultaneously bidirectionally translates between structural and functional brain connectivity across different atlases and processing variants ('flavors') via a common 128-dim L2-normalised latent representation. The Nature Methods 2025 publication's canonical model jointly encodes 15 flavors (3 atlases × {3 functional connectivity types + 2 structural tractography types}) and maps each to / from the shared latent.
+Architecture (per published recipe): per-flavor 256-dim PCA input transformation -> 256 -> 128 Linear encoder -> 128-dim L2-normalised latent -> 128 -> 256 Linear decoder -> inverse PCA to the destination flavor's full-dim connectivity space. v0 of this port ships the canonical bundle plus its 15-flavor PCA stack (separate ``krakencoder_pca_stack`` bundle that the ``KrakencoderPipeline`` co-loads).
+
+## Intended use
+
+Single-flavor targlatent encoder/decoder for SCifod2act_coco439_sift2volnorm. Pairs with krakencoder_pca.SCifod2act-coco439-sift2volnorm-pc256.1; designed to be merged into the canonical 15-flavor bundle via KrakencoderPipeline.with_extra_flavor().
+
+## Usage
+
+```python
+from ilex.models.krakencoder import Krakencoder
+model = Krakencoder.from_pretrained('ilex-hub/krakencoder.SCifod2act-coco439-sift2volnorm-targlatent.1')
+```
+
+## Authors
+
+Keith W. Jamison, Zijin Gu, Qinxin Wang, Ceren Tozlu, Mert R. Sabuncu, Amy Kuceyeski
+
+## Citation
+
+Jamison K.W., Gu Z., Wang Q., Tozlu C., Sabuncu M.R., Kuceyeski A. (2025). Krakencoder: a unified brain connectome translation and fusion tool. Nature Methods. DOI: 10.1038/s41592-025-02706-2.
+
+### References
+
+- Jamison K.W., Gu Z., Wang Q., Tozlu C., Sabuncu M.R., Kuceyeski A. (2025). Krakencoder: a unified brain connectome translation and fusion tool. Nature Methods. DOI: 10.1038/s41592-025-02706-2.
+- Preprint: bioRxiv 10.1101/2024.04.12.589274.
+- Upstream code: github.com/kjamison/krakencoder (model.py + fetch.py + per-flavor PCA transforms hosted on OSF: osf.io/dfp92).
+
+## License
+
+HF Hub license tag: `mit`
+
+**Effective terms:** MIT (copyright (c) 2024 Keith W. Jamison) on both the network code (github.com/kjamison/krakencoder) and the pretrained weights + per-flavor PCA transforms hosted on OSF (osf.io/dfp92). The ilex JAX / Equinox port code is separately licensed under Apache-2.0 / GPL-3.0.
+
+Upstream license reference: https://opensource.org/licenses/MIT
+
+### Copyright
+
+Network architecture, training code, and pretrained weights -- copyright (c) 2024 Keith W. Jamison; released under the MIT License. JAX / Equinox port code -- copyright (c) the ilex authors, released under the Apache-2.0 / GPL-3.0 dual license used by ilex itself.
+
+## Upstream source
+
+Original weights / reference implementation: https://github.com/kjamison/krakencoder
+
+## Provenance
+
+This artefact was produced by [ilex](https://github.com/hypercoil/ilex)'s
+save/load pipeline. The architecture is implemented in
+`ilex.models.krakencoder.Krakencoder` and the weights have been converted
+from their upstream format. See the upstream source above
+for the canonical reference.

config.json

@@ -0,0 +1,51 @@
+{
+  "_ilex": {
+    "architecture": "ilex.models.krakencoder.model.Krakencoder",
+    "constructor_kwargs": {
+      "input_size_list": [
+        256
+      ],
+      "latent_normalize": true,
+      "latentsize": 128,
+      "linear_include_bias": true
+    },
+    "flavor_name_list": [
+      "SCifod2act_coco439_sift2volnorm"
+    ],
+    "format": "ilex",
+    "framework_version": {
+      "equinox": "0.13.8",
+      "ilex": "0.0.0.dev0",
+      "jax": "0.10.0",
+      "jaxlib": "0.10.0",
+      "numpy": "2.4.4",
+      "safetensors": "0.7.0"
+    },
+    "has_state": false,
+    "origin": "ilex-native",
+    "weights_status": "bundled"
+  },
+  "authors": "Keith W. Jamison, Zijin Gu, Qinxin Wang, Ceren Tozlu, Mert R. Sabuncu, Amy Kuceyeski",
+  "copyright": "Network architecture, training code, and pretrained weights -- copyright (c) 2024 Keith W. Jamison; released under the MIT License. JAX / Equinox port code -- copyright (c) the ilex authors, released under the Apache-2.0 / GPL-3.0 dual license used by ilex itself.",
+  "data_type": "numpy",
+  "description": "Krakencoder (Jamison, Gu, Wang, Tozlu, Sabuncu, Kuceyeski, *Nature Methods* 2025), ported to JAX / Equinox from the upstream PyTorch release (github.com/kjamison/krakencoder). A linked autoencoder that simultaneously bidirectionally translates between structural and functional brain connectivity across different atlases and processing variants ('flavors') via a common 128-dim L2-normalised latent representation. The Nature Methods 2025 publication's canonical model jointly encodes 15 flavors (3 atlases \u00d7 {3 functional connectivity types + 2 structural tractography types}) and maps each to / from the shared latent.\nArchitecture (per published recipe): per-flavor 256-dim PCA input transformation -> 256 -> 128 Linear encoder -> 128-dim L2-normalised latent -> 128 -> 256 Linear decoder -> inverse PCA to the destination flavor's full-dim connectivity space. v0 of this port ships the canonical bundle plus its 15-flavor PCA stack (separate ``krakencoder_pca_stack`` bundle that the ``KrakencoderPipeline`` co-loads).",
+  "equinox_version": "0.13.8",
+  "ilex_version": "0.0.0.dev0",
+  "image_classes": "Brain connectivity matrices in any of the 15 supported flavors (per the canonical bundle's catalog row). Inputs are upper-triangular vectors -- one per flavor per subject. Atlas edge counts: FS86 = 3,655; Shen268 = 35,778; Coco439 = 96,141 (these are the full-dim sizes the per-flavor PCAs project from / inverse-project to).",
+  "intended_use": "Inference-time bidirectional translation between brain connectivity flavors via the shared latent representation. Pipeline:\n1. Caller provides input connectivity matrices (upper-triangular\n   vectors) in one or more of the 15 supported flavors.\n2. ``KrakencoderPipeline`` runs each per-flavor PCA forward\n   (full-dim -> 256-dim), encodes to the shared 128-dim\n   L2-normalised latent, then optionally decodes to one or more\n   target flavors and inverse-PCAs back to full dim.\n\nCommon use cases:\n* Translate input(s) in some flavors to predictions in others\n  (the Nature Methods paper's primary application).\n* Extract the 128-dim shared latent as a subject-level\n  representation for downstream prediction tasks (the\n  ``.encode()`` shortcut).\n* Fusion: average multiple input flavors' latents to improve\n  predictive identifiability.\n\nTrained on 700 subjects from the NIH Human Connectome Project; generalises to held-out HCP subjects and (per the paper) extends to non-HCP cohorts with some accuracy degradation.",
+  "jax_version": "0.10.0",
+  "network_data_format": {
+    "inputs": {},
+    "outputs": {}
+  },
+  "numpy_version": "2.4.4",
+  "pred_classes": "Outputs match the input shape: predicted connectivity matrices in the requested target flavor(s), as upper-triangular vectors at each flavor's atlas-determined edge count. The ``.encode()``-only shortcut emits a single (128,) L2-normalised latent per input subject.",
+  "references": [
+    "Jamison K.W., Gu Z., Wang Q., Tozlu C., Sabuncu M.R., Kuceyeski A. (2025). Krakencoder: a unified brain connectome translation and fusion tool. Nature Methods. DOI: 10.1038/s41592-025-02706-2.",
+    "Preprint: bioRxiv 10.1101/2024.04.12.589274.",
+    "Upstream code: github.com/kjamison/krakencoder (model.py + fetch.py + per-flavor PCA transforms hosted on OSF: osf.io/dfp92)."
+  ],
+  "schema": "https://github.com/Project-MONAI/MONAI-extra-test-data/releases/download/0.8.1/meta_schema_20240725.json",
+  "task": "brain-connectome translation (15 modality/atlas flavors via a shared latent space)",
+  "version": "0.0.0"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e596d9b366dfac36b67c50a18fe2ed0434f40f30e26d5b25670e9b00bdb672d
+size 264085