Push model using huggingface_hub.

README.md

@@ -1,98 +1,11 @@
 ---
-license: mit
 library_name: muscari
 tags:
-  - ecology
-  - biodiversity
-  - species-richness
-  - species-distribution
-  - vegetation
-  - Europe
-  - geospatial
-pretty_name: MuScaRi
-pipeline_tag: other
+- model_hub_mixin
+- pytorch_model_hub_mixin
 ---
 
-# Model Card for MuScaRi
-
-**MuScaRi** (Multi-Scale species Richness estimation, also named after the *Muscari* genus of perennial bulbous plants) is a deep learning model that estimates vascular plant species richness at arbitrary spatial scales from ecological survey data and environmental covariates.
-
-- **Repository:** https://github.com/vboussange/MuScaRi
-- **Paper:** [Multi-scale species richness estimation with deep learning](https://arxiv.org/abs/2507.06358)
-- **Training data:** [vboussange/muscari-data](https://huggingface.co/datasets/vboussange/muscari-data)
-- **Demo:** [![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/vboussange/MuScaRi/blob/master/muscari_demo.ipynb)
-
-## Model Description
-
-MuScaRi composes a fully connected feedforward neural network with a four-parameter Weibull rarefaction model. Given the area of a spatial unit and summary statistics of environmental covariates within it, the neural network predicts the parameters of the rarefaction curve, which in turn predicts expected species richness as a function of sampling effort. Evaluating the curve at infinite sampling effort yields total (asymptotic) species richness predictions.
-
-The pretrained model is an **ensemble of 5 members**, one per spatial cross-validation fold, trained on ~350k European vegetation plots from the European Vegetation Archive (EVA). Ensemble predictions are aggregated by arithmetic mean; standard deviations quantify prediction uncertainty.
-
-See the [paper](https://arxiv.org/abs/2507.06358) for full architecture details and benchmarks, and the [`muscari-data` dataset card](https://huggingface.co/datasets/vboussange/muscari-data) for the dataset used during training.
-
-## Quick Start
-
-```python
-from muscari import MuScaRiEnsemble
-from muscari.data_processing.utils_features import EnvironmentalFeatureDataset
-import pandas as pd
-
-model = MuScaRiEnsemble.from_pretrained("vboussange/muscari")
-print(f"Ensemble with {model.n_models} members")
-print("Required features:", model.feature_names)
-
-# Predict total species richness for a spatial unit
-# df must contain columns listed in model.feature_names
-df = pd.DataFrame([...])  # one row per spatial unit; see Colab demo for how to build it
-sr_mean = model.predict_mean_sr_tot(df)   # asymptotic richness
-sr_std  = model.get_std_sr_tot(df)        # ensemble uncertainty
-```
-
-For an end-to-end walkthrough, see the [Colab demo](https://colab.research.google.com/github/vboussange/MuScaRi/blob/master/muscari_demo.ipynb).
-
-## Inputs and Outputs
-
-**Inputs:**
-a `df: pandas.Dataframe` with the following columns (see [Colab demo](https://colab.research.google.com/github/vboussange/MuScaRi/blob/master/muscari_demo.ipynb) for more details)
-
-| Feature group | Columns | Description |
-|---|---|---|
-| Spatial unit area | `log_observed_area` | Log of sampling effort (m²); omit for asymptotic prediction |
-| Mean environmental conditions | mean of `bio1`, `bio12`, `sfcWind`, `pet`, `elevation` | Mean of CHELSA/EU-DEM variables within the spatial unit |
-| Environmental heterogeneity | std of `bio1`, `bio12`, `sfcWind`, `pet`, `elevation` | Std of CHELSA/EU-DEM variables within the spatial unit |
-
-**Outputs:**
-
-- `model.predict_mean_sr(df)`: expected species richness at a given sampling effort (interpolation mode)
-- `model.predict_mean_sr_tot(df)`: total species richness under asymptotic sampling effort (extrapolation mode)
-- `model.get_std_sr_tot(df)`: ensemble standard deviation of the above
-
-## Training Data and Evaluation
-
-Full performance tables are in the [paper](https://arxiv.org/abs/2507.06358).
-
-## Limitations
-
-- Trained on European vascular plants; performance outside Europe is untested.
-- Environmental predictors use a 1981-2010 climatological baseline.
-- Predictions are less reliable in data-sparse regions (e.g. parts of France, Spain, Scandinavia).
-
-## Citation
-
-```bibtex
-@misc{boussange2025muscari,
-  title         = {Multi-scale species richness estimation with deep learning},
-  author        = {Victor Boussange and Bert Wuyts and Philipp Brun and
-                   Johanna T. Malle and Gabriele Midolo and Jeanne Portier and
-                   Théophile Sanchez and Niklaus E. Zimmermann and
-                   Irena Axmanová and Helge Bruelheide and Milan Chytrý and
-                   Stephan Kambach and Zdeňka Lososová and Martin Večeřa and
-                   Idoia Biurrun and Klaus T. Ecker and Jonathan Lenoir and
-                   Jens-Christian Svenning and Dirk Nikolaus Karger},
-  year          = {2025},
-  eprint        = {2507.06358},
-  archivePrefix = {arXiv},
-  primaryClass  = {q-bio.PE},
-  url           = {https://arxiv.org/abs/2507.06358},
-}
-```
+This model has been pushed to the Hub using the [PytorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) integration:
+- Code: [More Information Needed]
+- Paper: [More Information Needed]
+- Docs: [More Information Needed]

config.json

@@ -0,0 +1,448 @@
+{
+  "feature_names": [
+    "bio1",
+    "pet_penman_mean",
+    "sfcWind_mean",
+    "bio12",
+    "std_bio1",
+    "std_pet_penman_mean",
+    "std_sfcWind_mean",
+    "std_bio12",
+    "elevation",
+    "std_elevation",
+    "log_sp_unit_area"
+  ],
+  "feature_scalers": [
+    {
+      "cls": "MinMaxScaler",
+      "data_max_": [
+        15.884361267089844,
+        19.202129364013672,
+        138.9114990234375,
+        10.632919311523438,
+        3517.006103515625,
+        6.485342502593994,
+        25.77168083190918,
+        2.8469786643981934,
+        1041.3431396484375,
+        2908.932861328125,
+        994.3865966796875,
+        27.631017684936523
+      ],
+      "data_min_": [
+        0.0,
+        -4.92936372756958,
+        33.97886657714844,
+        0.7180836200714111,
+        305.00048828125,
+        0.0,
+        0.012921497225761414,
+        0.002438093302771449,
+        0.04062916710972786,
+        -6.931583404541016,
+        0.0,
+        15.201862335205078
+      ],
+      "data_range_": [
+        15.884361267089844,
+        24.131492614746094,
+        104.93263244628906,
+        9.914835929870605,
+        3212.005615234375,
+        6.485342502593994,
+        25.758758544921875,
+        2.844540596008301,
+        1041.302490234375,
+        2915.864501953125,
+        994.3865966796875,
+        12.429155349731445
+      ],
+      "min_": [
+        0.0,
+        0.20427098870277405,
+        -0.3238160312175751,
+        -0.07242516428232193,
+        -0.0949564054608345,
+        0.0,
+        -0.000501635076943785,
+        -0.0008571131620556116,
+        -3.901763921021484e-05,
+        0.002377196680754423,
+        0.0,
+        -1.2230808734893799
+      ],
+      "n_features_in_": 12,
+      "scale_": [
+        0.06295499950647354,
+        0.04143962636590004,
+        0.009529924020171165,
+        0.10085895657539368,
+        0.0003113319689873606,
+        0.15419386327266693,
+        0.0388217456638813,
+        0.3515506088733673,
+        0.000960335717536509,
+        0.00034295147634111345,
+        0.0010056451428681612,
+        0.08045598864555359
+      ]
+    },
+    {
+      "cls": "MinMaxScaler",
+      "data_max_": [
+        15.843147277832031,
+        19.20088005065918,
+        139.17771911621094,
+        10.947614669799805,
+        3342.95263671875,
+        6.469884395599365,
+        24.22772216796875,
+        2.978724479675293,
+        1162.36328125,
+        3047.372314453125,
+        992.4478149414062,
+        27.6309871673584
+      ],
+      "data_min_": [
+        0.0,
+        -6.05493688583374,
+        34.17231369018555,
+        0.7988338470458984,
+        284.1519470214844,
+        0.0,
+        0.010869947262108326,
+        0.004524758085608482,
+        0.013255664147436619,
+        -6.974416255950928,
+        0.0,
+        15.20180606842041
+      ],
+      "data_range_": [
+        15.843147277832031,
+        25.255817413330078,
+        105.00540161132812,
+        10.148780822753906,
+        3058.80078125,
+        6.469884395599365,
+        24.21685218811035,
+        2.9741997718811035,
+        1162.3499755859375,
+        3054.3466796875,
+        992.4478149414062,
+        12.429181098937988
+      ],
+      "min_": [
+        0.0,
+        0.23974423110485077,
+        -0.3254338800907135,
+        -0.07871229946613312,
+        -0.09289652109146118,
+        0.0,
+        -0.00044885880197398365,
+        -0.0015213362639769912,
+        -1.1404194083297625e-05,
+        0.002283439738675952,
+        0.0,
+        -1.2230738401412964
+      ],
+      "n_features_in_": 12,
+      "scale_": [
+        0.06311877071857452,
+        0.03959483653306961,
+        0.009523320011794567,
+        0.0985340029001236,
+        0.00032692551030777395,
+        0.15456226468086243,
+        0.04129355773329735,
+        0.33622488379478455,
+        0.0008603261085227132,
+        0.0003274022601544857,
+        0.0010076096514239907,
+        0.08045582473278046
+      ]
+    },
+    {
+      "cls": "MinMaxScaler",
+      "data_max_": [
+        15.886514663696289,
+        19.208925247192383,
+        142.9532928466797,
+        10.431619644165039,
+        3432.67919921875,
+        6.512694835662842,
+        25.317461013793945,
+        2.8370885848999023,
+        1105.497314453125,
+        2948.165283203125,
+        996.5736694335938,
+        27.631013870239258
+      ],
+      "data_min_": [
+        0.0,
+        -5.62136173248291,
+        33.247920989990234,
+        0.7645986676216125,
+        306.2086181640625,
+        0.0,
+        0.010618088766932487,
+        0.0022086116950958967,
+        0.013255664147436619,
+        -7.108325481414795,
+        0.0,
+        15.201883316040039
+      ],
+      "data_range_": [
+        15.886514663696289,
+        24.83028793334961,
+        109.70536804199219,
+        9.667020797729492,
+        3126.470703125,
+        6.512694835662842,
+        25.306842803955078,
+        2.8348798751831055,
+        1105.4840087890625,
+        2955.273681640625,
+        996.5736694335938,
+        12.429130554199219
+      ],
+      "min_": [
+        0.0,
+        0.22639131546020508,
+        -0.30306559801101685,
+        -0.07909351587295532,
+        -0.09794066101312637,
+        0.0,
+        -0.00041957382927648723,
+        -0.0007790847448632121,
+        -1.1990823622909375e-05,
+        0.0024053019005805254,
+        0.0,
+        -1.2230850458145142
+      ],
+      "n_features_in_": 12,
+      "scale_": [
+        0.06294646859169006,
+        0.04027339443564415,
+        0.009115324355661869,
+        0.10344448685646057,
+        0.0003198494669049978,
+        0.1535462737083435,
+        0.03951500356197357,
+        0.3527486324310303,
+        0.0009045811602845788,
+        0.00033837812952697277,
+        0.0010034381411969662,
+        0.08045615255832672
+      ]
+    },
+    {
+      "cls": "MinMaxScaler",
+      "data_max_": [
+        15.892108917236328,
+        19.225000381469727,
+        138.10855102539062,
+        10.26804256439209,
+        3565.4345703125,
+        6.537107944488525,
+        25.24667739868164,
+        2.9901700019836426,
+        1012.9049682617188,
+        2916.74609375,
+        998.977294921875,
+        27.630943298339844
+      ],
+      "data_min_": [
+        0.0,
+        -5.0402984619140625,
+        34.54750061035156,
+        0.7479619979858398,
+        286.7718811035156,
+        0.0,
+        0.010618088766932487,
+        0.0032483888790011406,
+        0.013255664147436619,
+        -7.108325481414795,
+        0.0,
+        15.201833724975586
+      ],
+      "data_range_": [
+        15.892108917236328,
+        24.26529884338379,
+        103.56105041503906,
+        9.52008056640625,
+        3278.66259765625,
+        6.537107944488525,
+        25.236059188842773,
+        2.986921548843384,
+        1012.8917236328125,
+        2923.8544921875,
+        998.977294921875,
+        12.429109573364258
+      ],
+      "min_": [
+        0.0,
+        0.20771631598472595,
+        -0.33359548449516296,
+        -0.078566774725914,
+        -0.08746611326932907,
+        0.0,
+        -0.0004207506717648357,
+        -0.0010875373845919967,
+        -1.3086951184959617e-05,
+        0.0024311488959938288,
+        0.0,
+        -1.2230831384658813
+      ],
+      "n_features_in_": 12,
+      "scale_": [
+        0.06292431056499481,
+        0.04121111333370209,
+        0.009656139649450779,
+        0.10504113137722015,
+        0.00030500240973196924,
+        0.15297284722328186,
+        0.03962583839893341,
+        0.3347928524017334,
+        0.0009872723603621125,
+        0.0003420142747927457,
+        0.0010010238038375974,
+        0.08045628666877747
+      ]
+    },
+    {
+      "cls": "MinMaxScaler",
+      "data_max_": [
+        15.827322959899902,
+        19.24164390563965,
+        135.46925354003906,
+        10.365579605102539,
+        3569.140869140625,
+        6.496120929718018,
+        25.69937515258789,
+        2.7845206260681152,
+        1086.6605224609375,
+        2914.540771484375,
+        967.4793090820312,
+        27.63096809387207
+      ],
+      "data_min_": [
+        0.0,
+        -4.978337287902832,
+        33.97886657714844,
+        0.7802224159240723,
+        309.584716796875,
+        0.0,
+        0.010618088766932487,
+        0.0027105531189590693,
+        0.09874647110700607,
+        -7.108325481414795,
+        0.0,
+        15.201906204223633
+      ],
+      "data_range_": [
+        15.827322959899902,
+        24.219982147216797,
+        101.49038696289062,
+        9.585357666015625,
+        3259.55615234375,
+        6.496120929718018,
+        25.688756942749023,
+        2.7818100452423096,
+        1086.561767578125,
+        2921.649169921875,
+        967.4793090820312,
+        12.429061889648438
+      ],
+      "min_": [
+        0.0,
+        0.20554670691490173,
+        -0.3347988724708557,
+        -0.08139731734991074,
+        -0.09497756510972977,
+        0.0,
+        -0.00041333603439852595,
+        -0.0009743847185745835,
+        -9.087975922739133e-05,
+        0.0024329840671271086,
+        0.0,
+        -1.2230935096740723
+      ],
+      "n_features_in_": 12,
+      "scale_": [
+        0.06318187713623047,
+        0.04128822311758995,
+        0.009853149764239788,
+        0.1043257862329483,
+        0.00030679022893309593,
+        0.15393802523612976,
+        0.03892753645777702,
+        0.3594781756401062,
+        0.000920334248803556,
+        0.0003422724548727274,
+        0.0010336138075217605,
+        0.08045659214258194
+      ]
+    }
+  ],
+  "layer_sizes": [
+    128,
+    512,
+    2048,
+    2048,
+    512,
+    128
+  ],
+  "n_models": 5,
+  "target_scalers": [
+    {
+      "cls": "MaxAbsScaler",
+      "max_abs_": [
+        4755.0
+      ],
+      "n_features_in_": 1,
+      "scale_": [
+        4755.0
+      ]
+    },
+    {
+      "cls": "MaxAbsScaler",
+      "max_abs_": [
+        4772.0
+      ],
+      "n_features_in_": 1,
+      "scale_": [
+        4772.0
+      ]
+    },
+    {
+      "cls": "MaxAbsScaler",
+      "max_abs_": [
+        4864.0
+      ],
+      "n_features_in_": 1,
+      "scale_": [
+        4864.0
+      ]
+    },
+    {
+      "cls": "MaxAbsScaler",
+      "max_abs_": [
+        4725.0
+      ],
+      "n_features_in_": 1,
+      "scale_": [
+        4725.0
+      ]
+    },
+    {
+      "cls": "MaxAbsScaler",
+      "max_abs_": [
+        4701.0
+      ],
+      "n_features_in_": 1,
+      "scale_": [
+        4701.0
+      ]
+    }
+  ]
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be3792b5290693a8f9102ac26f5107c0ce12398eba06a9a6355116c86ee06419
+size 129054112