add model

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+anvil_training/model.pth filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,3 +1,124 @@
 ---
 license: apache-2.0
+language:
+- en
+tags:
+- chemistry
+- drug-discovery
+- admet
+- multitask-learning
+- openadmet
 ---
+
+This is our baseline Caco-2 permeability/LogD/PPB model. It is a **multitask CheMeleon model** trained to predict the following endpoints:
+
+- Caco-2 Permeability Papp A->B
+- Caco-2 Permeability Papp B->A
+- LogD
+- MPPB
+- HPPB
+
+## Pre-requisites
+We *highly* recommend you have the Anvil framework from `openadmet-models` installed in an environment (called `openadmet-models`) for ease of use and full utilization of OpenADMET's models. For full documentation, visit our website [here](https://docs.openadmet.org/en/latest/). If you'd like to see some more examples on how to use Anvil, see our demos [here](https://demos.openadmet.org/en/latest/).
+
+### Installation of `openadmet-models`
+
+#### With conda
+
+You can install openadmet-models via our GitHub package. If you want the latest development version, clone the repository and install in editable mode:
+
+```
+git clone git@github.com:OpenADMET/openadmet-models.git
+```
+
+Set up an environment using the provided files in devtools/conda-envs.
+```
+cd openadmet-models/
+conda env create -f devtools/conda-envs/openadmet-models.yaml
+conda activate openadmet-models
+pip install -e .
+```
+If you want to use GPU acceleration, ensure you have the appropriate CUDA toolkit installed and use the openadmet-models-gpu.yaml file instead:
+```
+conda env create -f devtools/conda-envs/openadmet-models-gpu.yaml
+conda activate openadmet-models
+pip install -e .
+```
+#### With Docker
+Alternatively, you can also use Docker to spin up a containerized pre-installed environment to run `openadmet-models`. Just be sure you are mounting the correct folder (`./permeability-logd-ppb-chemeleon-baseline`) where you've downloaded the model.
+
+If you're using a gpu, run:
+```
+docker run -it --user=root --rm  \
+	-v ./permeability-logd-ppb-chemeleon-baseline:/home/mambauser/model:rw \
+	--runtime=nvidia \
+	--gpus \
+	all ghcr.io/openadmet/openadmet-models:main
+```
+Otherwise, for cpu only:
+```
+docker run -it --user=root --rm  \
+	-v ./permeability-logd-ppb-chemeleon-baseline:/home/mambauser/model:rw \
+	all ghcr.io/openadmet/openadmet-models:main
+```
+
+**IMPORTANT NOTE** You will also need `git lfs` installed.
+
+## Downloading the model
+
+1. After installing Anvil, clone the model repo:
+```
+git clone https://huggingface.co/openadmet/permeability-logd-ppb-chemeleon-baseline/
+```
+2. Change to the repo directory. Ensure you have `git lfs` installed for the repo and get the large model files:
+```
+git lfs install
+git lfs pull
+```
+3. You are now ready to use the model!
+
+## Using the model
+
+**IMPORTANT NOTE** This model predicts $\log_{10}(P_{app})$ values (on $\log_{10}(cm/s)$). To get $P_{app}$ values in $10^{-6}cm/s$, simply backtransform:
+$$
+P_{app} = 10^{\hat{y}} * 10^{6}
+$$
+
+Where $\hat{y}$ is our model prediction. 
+
+We will use this model for inference, to predict endpoint values for a set of molecular compounds unseen to the model. For demonstration purposes, we will be using a small-molecule set from our recent [OpenADMET-ExpansionRx challenge](https://huggingface.co/spaces/openadmet/OpenADMET-ExpansionRx-Challenge), provided in the file `expansion_data_inference.csv`.
+You can do this either **inside the docker container** as per the instructions above, or if you have installed openadmet-models on your own computer, you can use the appropriate environment.
+
+The generic command to run our inference pipeline is:
+```bash
+openadmet predict \
+	--input-path <the path to the data to predict on> \
+	--input-col <the column of the data to predict on, often SMILES> \
+	--model-dir <the anvil_training directory of the model to predict with> \
+	--output-csv <the path to an output CSV to save the predictions to> \
+	--accelerator <whether to use gpu or cpu, defaults to gpu>
+```
+You can run this directly in your command line, OR you can use the bash script we've provided, `run_model_inference.sh`.
+
+For our working example, this command becomes:
+```bash
+openadmet predict \
+	--input-path expansion_data_inference.csv \
+	--input-col SMILES \
+	--model-dir anvil_training/ \
+	--output-csv predictions.csv \
+	--accelerator cpu
+```
+You can easily substitute your own set of compounds, simply modify the `--input-path` and `--input-col` arguments for your specific dataset.
+
+In our example, this outputs a file called `predictions.csv` which includes endpoint-specific prediction columns (as `OADMET_PRED_chemprop_{}`) for:
+
+- `caco2_atob_LogPapp`
+- `caco2_btoa_LogPapp`
+- `logD`
+- `mppb_LogUnbound`
+- `hppb_LogUnbound`
+
+In this case, `OADMET_STD_chemprop_{}` columns are empty because uncertainty cannot be estimated unless running inference on an ensemble of models. See how to set this option [here](https://demos.openadmet.org/en/latest/demos/04_Ensemble_Model_Training/04_Ensemble_Model_Training_Active_Learning.html).
+
+**IMPORTANT NOTE** If you'd like other examples for how to use our Anvil framework, checkout our demos [here](https://demos.openadmet.org/en/latest/).

anvil_training/anvil_recipe.yaml

@@ -0,0 +1,93 @@
+data:
+  anvil_dir: file:///home/ec2-user/permeability_models/multitask/all_endpoints_model
+  cat_entry: null
+  dropna: false
+  input_col: OPENADMET_CANONICAL_SMILES
+  resource: ../../ChEMBL35_Caco2_permeability_multitask_atob_btoa_logD_mppb_rescaled.parquet
+  target_cols:
+  - logD
+  - caco2_atob_LogPapp
+  - caco2_btoa_LogPapp
+  - mppb_LogUnbound
+  - hppb_LogUnbound
+  test_resource: null
+  train_resource: null
+  type: intake
+  val_resource: null
+metadata:
+  authors: Maria A. Castellanos
+  biotargets:
+  - Caco-2
+  build_number: 0
+  description: basic regression using a ChemProp multitask task model
+  driver: pytorch
+  email: maria.castellanos@omsf.io
+  name: chemprop_pchembl
+  tag: chemprop
+  tags:
+  - openadmet
+  - example
+  - chemprop
+  version: v1
+procedure:
+  ensemble: null
+  feat:
+    params:
+      batch_size: 128
+      n_jobs: 4
+      normalize_targets: true
+    type: ChemPropFeaturizer
+  model:
+    freeze_weights: null
+    param_path: null
+    params:
+      batch_norm: false
+      dropout: 0.25
+      ffn_hidden_dim: 512
+      ffn_hidden_num_layers: 3
+      ffn_lr: 1e-3
+      ffn_weight_decay: 1e-4
+      from_chemeleon: true
+      mpnn_lr: 1e-4
+      mpnn_weight_decay: 0
+      n_tasks: 5
+      reduce_lr_factor: 0.5
+      reduce_lr_patience: 5
+      scheduler: plateau
+    serial_path: null
+    type: ChemPropModel
+  split:
+    params:
+      random_state: 42
+      test_size: 0.0
+      train_size: 1.0
+      val_size: 0.0
+    type: ShuffleSplitter
+  train:
+    params:
+      accelerator: gpu
+      early_stopping: true
+      early_stopping_min_delta: 0.001
+      early_stopping_mode: min
+      early_stopping_patience: 20
+      gradient_clip_val: 0.5
+      max_epochs: 200
+      monitor_metric: val_loss
+      use_wandb: false
+      wandb_project: demos
+    type: LightningTrainer
+  transform: null
+report:
+  eval:
+  - params: {}
+    type: RegressionMetrics
+  - params:
+      axes_labels:
+      - True LogPapp
+      - Predicted LogPapp
+      n_repeats: 5
+      n_splits: 5
+      pXC50: true
+      random_state: 42
+      title: Multitask True vs Predicted LogPapp on test set
+    type: PytorchLightningRepeatedKFoldCrossValidation

anvil_training/logs/model/version_0/hparams.yaml

@@ -0,0 +1,5 @@
+batch_norm: false
+final_lr: 1.0e-05
+init_lr: 0.0001
+max_lr: 0.001
+warmup_epochs: 2

anvil_training/logs/model/version_0/metrics.csv

@@ -0,0 +1,799 @@
+epoch,step,train_loss_epoch,train_loss_step
+0,49,,0.9781196713447571
+0,99,,0.5188053250312805
+0,149,,0.4172388017177582
+0,199,,0.5668007135391235
+0,232,0.5967047810554504,
+1,249,,0.40170538425445557
+1,299,,0.31435340642929077
+1,349,,0.31079158186912537
+1,399,,0.5218468308448792
+1,449,,0.6680266261100769
+1,465,0.3945891261100769,
+2,499,,0.32706037163734436
+2,549,,0.23362305760383606
+2,599,,0.5701060891151428
+2,649,,0.4605569839477539
+2,698,0.31467100977897644,
+3,699,,0.2150094211101532
+3,749,,0.43613582849502563
+3,799,,0.4241478145122528
+3,849,,0.24332112073898315
+3,899,,0.17995944619178772
+3,931,0.26194003224372864,
+4,949,,0.1358797401189804
+4,999,,0.17095600068569183
+4,1049,,0.1124139279127121
+4,1099,,0.17797376215457916
+4,1149,,0.1393575817346573
+4,1164,0.2246982306241989,
+5,1199,,0.15030537545681
+5,1249,,0.14297890663146973
+5,1299,,0.15206843614578247
+5,1349,,0.13339249789714813
+5,1397,0.191450834274292,
+6,1399,,0.061842773109674454
+6,1449,,0.13477012515068054
+6,1499,,0.0641542449593544
+6,1549,,0.09694936126470566
+6,1599,,0.09821660071611404
+6,1630,0.16744816303253174,
+7,1649,,0.1417410969734192
+7,1699,,0.1403740793466568
+7,1749,,0.06036369130015373
+7,1799,,0.25044381618499756
+7,1849,,0.3270062506198883
+7,1863,0.1461438238620758,
+8,1899,,0.08017916977405548
+8,1949,,0.2208152860403061
+8,1999,,0.07828712463378906
+8,2049,,0.1835923045873642
+8,2096,0.1315426081418991,
+9,2099,,0.17264235019683838
+9,2149,,0.0769258588552475
+9,2199,,0.08432567119598389
+9,2249,,0.06482519209384918
+9,2299,,0.048316650092601776
+9,2329,0.12243183702230453,
+10,2349,,0.0460486114025116
+10,2399,,0.114642433822155
+10,2449,,0.14394474029541016
+10,2499,,0.09048590064048767
+10,2549,,0.10425841808319092
+10,2562,0.10534672439098358,
+11,2599,,0.05577409267425537
+11,2649,,0.06805107742547989
+11,2699,,0.06150142103433609
+11,2749,,0.33150267601013184
+11,2795,0.10354717075824738,
+12,2799,,0.07958871871232986
+12,2849,,0.1425044983625412
+12,2899,,0.06924302130937576
+12,2949,,0.04331598058342934
+12,2999,,0.07445185631513596
+12,3028,0.10292667895555496,
+13,3049,,0.07336984574794769
+13,3099,,0.08479808270931244
+13,3149,,0.05507899448275566
+13,3199,,0.05481494590640068
+13,3249,,0.08469651639461517
+13,3261,0.089226134121418,
+14,3299,,0.04025832936167717
+14,3349,,0.05584246292710304
+14,3399,,0.06666907668113708
+14,3449,,0.07012591511011124
+14,3494,0.07838718593120575,
+15,3499,,0.056400880217552185
+15,3549,,0.049936987459659576
+15,3599,,0.07748711109161377
+15,3649,,0.029488475993275642
+15,3699,,0.05908187851309776
+15,3727,0.08028434962034225,
+16,3749,,0.07860235124826431
+16,3799,,0.0880601704120636
+16,3849,,0.04102602228522301
+16,3899,,0.04515751451253891
+16,3949,,0.19417811930179596
+16,3960,0.083133265376091,
+17,3999,,0.11461348831653595
+17,4049,,0.035498738288879395
+17,4099,,0.08330532908439636
+17,4149,,0.08714041858911514
+17,4193,0.06922087073326111,
+18,4199,,0.03952097147703171
+18,4249,,0.06284266710281372
+18,4299,,0.050608083605766296
+18,4349,,0.022302500903606415
+18,4399,,0.14734108746051788
+18,4426,0.0700310543179512,
+19,4449,,0.03744671121239662
+19,4499,,0.06919052451848984
+19,4549,,0.038604214787483215
+19,4599,,0.051261648535728455
+19,4649,,0.0611330009996891
+19,4659,0.06360835582017899,
+20,4699,,0.09267693758010864
+20,4749,,0.06001728028059006
+20,4799,,0.0814255028963089
+20,4849,,0.13632343709468842
+20,4892,0.06337371468544006,
+21,4899,,0.0676366463303566
+21,4949,,0.06169576570391655
+21,4999,,0.09422094374895096
+21,5049,,0.023442301899194717
+21,5099,,0.037745408713817596
+21,5125,0.06597426533699036,
+22,5149,,0.16991564631462097
+22,5199,,0.04800787195563316
+22,5249,,0.05044173449277878
+22,5299,,0.02857021801173687
+22,5349,,0.04038986563682556
+22,5358,0.07110031694173813,
+23,5399,,0.05177393555641174
+23,5449,,0.05454722419381142
+23,5499,,0.06335316598415375
+23,5549,,0.0391240194439888
+23,5591,0.050620272755622864,
+24,5599,,0.02705218642950058
+24,5649,,0.048712193965911865
+24,5699,,0.04236499220132828
+24,5749,,0.025016091763973236
+24,5799,,0.2223322093486786
+24,5824,0.05181926116347313,
+25,5849,,0.0888863205909729
+25,5899,,0.05919940397143364
+25,5949,,0.07385910302400589
+25,5999,,0.07656605541706085
+25,6049,,0.13181981444358826
+25,6057,0.0656149610877037,
+26,6099,,0.05413954704999924
+26,6149,,0.05523274093866348
+26,6199,,0.038041986525058746
+26,6249,,0.018460527062416077
+26,6290,0.055361028760671616,
+27,6299,,0.06857875734567642
+27,6349,,0.023595184087753296
+27,6399,,0.04772178828716278
+27,6449,,0.06056066229939461
+27,6499,,0.06419448554515839
+27,6523,0.0560297928750515,
+28,6549,,0.03243432939052582
+28,6599,,0.02729768492281437
+28,6649,,0.03501734137535095
+28,6699,,0.03558112308382988
+28,6749,,0.046969976276159286
+28,6756,0.040195345878601074,
+29,6799,,0.048745106905698776
+29,6849,,0.06518510729074478
+29,6899,,0.03547127917408943
+29,6949,,0.05089491233229637
+29,6989,0.05161628499627113,
+30,6999,,0.036542363464832306
+30,7049,,0.05947002023458481
+30,7099,,0.05110417678952217
+30,7149,,0.049922600388526917
+30,7199,,0.02692374773323536
+30,7222,0.048779379576444626,
+31,7249,,0.03379150852560997
+31,7299,,0.07708773016929626
+31,7349,,0.04396437481045723
+31,7399,,0.06993021816015244
+31,7449,,0.0570831373333931
+31,7455,0.05120982602238655,
+32,7499,,0.03399725630879402
+32,7549,,0.14329299330711365
+32,7599,,0.02264322340488434
+32,7649,,0.039611149579286575
+32,7688,0.05079389363527298,
+33,7699,,0.09805820137262344
+33,7749,,0.03994247689843178
+33,7799,,0.04653069004416466
+33,7849,,0.054189737886190414
+33,7899,,0.0407494455575943
+33,7921,0.052060578018426895,
+34,7949,,0.02590177021920681
+34,7999,,0.04000856727361679
+34,8049,,0.04667581245303154
+34,8099,,0.03172371909022331
+34,8149,,0.03767998516559601
+34,8154,0.039249662309885025,
+35,8199,,0.07415590435266495
+35,8249,,0.043197132647037506
+35,8299,,0.04284992069005966
+35,8349,,0.0719987154006958
+35,8387,0.04648414999246597,
+36,8399,,0.04710324853658676
+36,8449,,0.02300715073943138
+36,8499,,0.027295563369989395
+36,8549,,0.037202540785074234
+36,8599,,0.031333908438682556
+36,8620,0.0447365865111351,
+37,8649,,0.03890945389866829
+37,8699,,0.03619258478283882
+37,8749,,0.019218752160668373
+37,8799,,0.025540076196193695
+37,8849,,0.04414200782775879
+37,8853,0.0368005707859993,
+38,8899,,0.024949198588728905
+38,8949,,0.024182504042983055
+38,8999,,0.04157062992453575
+38,9049,,0.03511962294578552
+38,9086,0.038144517689943314,
+39,9099,,0.034238122403621674
+39,9149,,0.07992681860923767
+39,9199,,0.035583410412073135
+39,9249,,0.031946223229169846
+39,9299,,0.039652250707149506
+39,9319,0.04573826119303703,
+40,9349,,0.03342490270733833
+40,9399,,0.03202459588646889
+40,9449,,0.016030456870794296
+40,9499,,0.02725359797477722
+40,9549,,0.04667072743177414
+40,9552,0.03469839319586754,
+41,9599,,0.02457886002957821
+41,9649,,0.029376346617937088
+41,9699,,0.025892267003655434
+41,9749,,0.07114774733781815
+41,9785,0.032746974378824234,
+42,9799,,0.04335075989365578
+42,9849,,0.03730068728327751
+42,9899,,0.0644807368516922
+42,9949,,0.029070377349853516
+42,9999,,0.0665367990732193
+42,10018,0.043127235025167465,
+43,10049,,0.022104596719145775
+43,10099,,0.04854988306760788
+43,10149,,0.052581265568733215
+43,10199,,0.0481957346200943
+43,10249,,0.016892345622181892
+43,10251,0.04314756393432617,
+44,10299,,0.02020653523504734
+44,10349,,0.04440908133983612
+44,10399,,0.02124502882361412
+44,10449,,0.06817551702260971
+44,10484,0.04067850485444069,
+45,10499,,0.016907187178730965
+45,10549,,0.03728348761796951
+45,10599,,0.06583327800035477
+45,10649,,0.038291964679956436
+45,10699,,0.030495375394821167
+45,10717,0.03282175958156586,
+46,10749,,0.02048608474433422
+46,10799,,0.024171175435185432
+46,10849,,0.029224112629890442
+46,10899,,0.03343479335308075
+46,10949,,0.02478986606001854
+46,10950,0.037528570741415024,
+47,10999,,0.029862411320209503
+47,11049,,0.03713463991880417
+47,11099,,0.03420879319310188
+47,11149,,0.07011330872774124
+47,11183,0.041649945080280304,
+48,11199,,0.027073826640844345
+48,11249,,0.025814339518547058
+48,11299,,0.04928593710064888
+48,11349,,0.03412342816591263
+48,11399,,0.036903250962495804
+48,11416,0.038343049585819244,
+49,11449,,0.03057154454290867
+49,11499,,0.031167030334472656
+49,11549,,0.013768763281404972
+49,11599,,0.027419738471508026
+49,11649,,0.02381688356399536
+49,11649,0.026815785095095634,
+50,11699,,0.04308406263589859
+50,11749,,0.014739257283508778
+50,11799,,0.04027842730283737
+50,11849,,0.021697254851460457
+50,11882,0.022889846935868263,
+51,11899,,0.009665418416261673
+51,11949,,0.015529194846749306
+51,11999,,0.03188738971948624
+51,12049,,0.03507628291845322
+51,12099,,0.020320162177085876
+51,12115,0.02245517261326313,
+52,12149,,0.020910289138555527
+52,12199,,0.014244514517486095
+52,12249,,0.02076144516468048
+52,12299,,0.04021155461668968
+52,12348,0.02100306749343872,
+53,12349,,0.015516445972025394
+53,12399,,0.029440879821777344
+53,12449,,0.026881836354732513
+53,12499,,0.04428185522556305
+53,12549,,0.013184783980250359
+53,12581,0.02054700255393982,
+54,12599,,0.014740287326276302
+54,12649,,0.014811485074460506
+54,12699,,0.050230178982019424
+54,12749,,0.01579219475388527
+54,12799,,0.013154199346899986
+54,12814,0.019415931776165962,
+55,12849,,0.01759403944015503
+55,12899,,0.022916141897439957
+55,12949,,0.011409372091293335
+55,12999,,0.01176859438419342
+55,13047,0.019173450767993927,
+56,13049,,0.01411499921232462
+56,13099,,0.015891948714852333
+56,13149,,0.014531861990690231
+56,13199,,0.020477071404457092
+56,13249,,0.019630191847682
+56,13280,0.02156912162899971,
+57,13299,,0.030056564137339592
+57,13349,,0.01568136364221573
+57,13399,,0.011954088695347309
+57,13449,,0.013162550516426563
+57,13499,,0.07792185246944427
+57,13513,0.02316400781273842,
+58,13549,,0.025784768164157867
+58,13599,,0.024627553299069405
+58,13649,,0.030907295644283295
+58,13699,,0.02679361216723919
+58,13746,0.02366592362523079,
+59,13749,,0.01159858237951994
+59,13799,,0.033839330077171326
+59,13849,,0.03137169033288956
+59,13899,,0.009843239560723305
+59,13949,,0.009268684312701225
+59,13979,0.023474013432860374,
+60,13999,,0.03938756510615349
+60,14049,,0.02775568515062332
+60,14099,,0.02392936870455742
+60,14149,,0.01927240937948227
+60,14199,,0.017045823857188225
+60,14212,0.021403346210718155,
+61,14249,,0.009746083058416843
+61,14299,,0.01225332636386156
+61,14349,,0.014065410010516644
+61,14399,,0.02647019736468792
+61,14445,0.019574718549847603,
+62,14449,,0.01705792360007763
+62,14499,,0.04200773313641548
+62,14549,,0.016878720372915268
+62,14599,,0.020176678895950317
+62,14649,,0.009628033265471458
+62,14678,0.024437952786684036,
+63,14699,,0.012443630956113338
+63,14749,,0.015057109296321869
+63,14799,,0.01606735773384571
+63,14849,,0.013969162479043007
+63,14899,,0.02002536505460739
+63,14911,0.020444611087441444,
+64,14949,,0.03022288717329502
+64,14999,,0.01323009468615055
+64,15049,,0.02545197680592537
+64,15099,,0.010245569981634617
+64,15144,0.01773940585553646,
+65,15149,,0.013383843004703522
+65,15199,,0.011288399808108807
+65,15249,,0.01123131811618805
+65,15299,,0.011422884650528431
+65,15349,,0.010151165537536144
+65,15377,0.01607241854071617,
+66,15399,,0.017162185162305832
+66,15449,,0.012982919812202454
+66,15499,,0.009785326197743416
+66,15549,,0.008285196498036385
+66,15599,,0.023085830733180046
+66,15610,0.01536955963820219,
+67,15649,,0.03951169550418854
+67,15699,,0.008769160136580467
+67,15749,,0.019500715658068657
+67,15799,,0.010686660185456276
+67,15843,0.014380306005477905,
+68,15849,,0.011538452468812466
+68,15899,,0.016681242734193802
+68,15949,,0.010189036838710308
+68,15999,,0.00833972729742527
+68,16049,,0.011182776652276516
+68,16076,0.013873722404241562,
+69,16099,,0.012912469916045666
+69,16149,,0.011790958233177662
+69,16199,,0.009379038587212563
+69,16249,,0.016848187893629074
+69,16299,,0.011755463667213917
+69,16309,0.014158600009977818,
+70,16349,,0.030424807220697403
+70,16399,,0.014991297386586666
+70,16449,,0.01172870397567749
+70,16499,,0.01876993663609028
+70,16542,0.013123802840709686,
+71,16549,,0.015601775608956814
+71,16599,,0.010485561564564705
+71,16649,,0.009183624759316444
+71,16699,,0.008191649802029133
+71,16749,,0.012665375135838985
+71,16775,0.013269945047795773,
+72,16799,,0.016645630821585655
+72,16849,,0.013750598765909672
+72,16899,,0.012924432754516602
+72,16949,,0.007964813150465488
+72,16999,,0.010726666077971458
+72,17008,0.012400693260133266,
+73,17049,,0.014438719488680363
+73,17099,,0.012546678073704243
+73,17149,,0.01671169139444828
+73,17199,,0.012266173027455807
+73,17241,0.013181381858885288,
+74,17249,,0.009332988411188126
+74,17299,,0.009313647635281086
+74,17349,,0.006196092814207077
+74,17399,,0.007162331137806177
+74,17449,,0.022599704563617706
+74,17474,0.012488004751503468,
+75,17499,,0.010622385889291763
+75,17549,,0.00992902647703886
+75,17599,,0.01027025654911995
+75,17649,,0.011277811601758003
+75,17699,,0.010832429863512516
+75,17707,0.012438989244401455,
+76,17749,,0.012987510301172733
+76,17799,,0.008428554981946945
+76,17849,,0.009375380352139473
+76,17899,,0.007473239209502935
+76,17940,0.012325131334364414,
+77,17949,,0.014484778977930546
+77,17999,,0.0075533525086939335
+77,18049,,0.008731091395020485
+77,18099,,0.011290667578577995
+77,18149,,0.013907559216022491
+77,18173,0.011834160424768925,
+78,18199,,0.009607281535863876
+78,18249,,0.009125184267759323
+78,18299,,0.009831167757511139
+78,18349,,0.01084428746253252
+78,18399,,0.08651471883058548
+78,18406,0.012310071848332882,
+79,18449,,0.00999101810157299
+79,18499,,0.02118723653256893
+79,18549,,0.008504064753651619
+79,18599,,0.014103724621236324
+79,18639,0.012438567355275154,
+80,18649,,0.012329409830272198
+80,18699,,0.02373524196445942
+80,18749,,0.007249414920806885
+80,18799,,0.01257007010281086
+80,18849,,0.009655717760324478
+80,18872,0.012100978754460812,
+81,18899,,0.008723409846425056
+81,18949,,0.008556880056858063
+81,18999,,0.0121050039306283
+81,19049,,0.01161923911422491
+81,19099,,0.011274706572294235
+81,19105,0.012260306626558304,
+82,19149,,0.009180237539112568
+82,19199,,0.015189701691269875
+82,19249,,0.01104743778705597
+82,19299,,0.008228017948567867
+82,19338,0.012209893204271793,
+83,19349,,0.013740824535489082
+83,19399,,0.009296142496168613
+83,19449,,0.012390985153615475
+83,19499,,0.02199133299291134
+83,19549,,0.007197687402367592
+83,19571,0.012731318362057209,
+84,19599,,0.01804499328136444
+84,19649,,0.025122037157416344
+84,19699,,0.00985637865960598
+84,19749,,0.008265969343483448
+84,19799,,0.01638331264257431
+84,19804,0.014814550057053566,
+85,19849,,0.01603962481021881
+85,19899,,0.008784090168774128
+85,19949,,0.009367906488478184
+85,19999,,0.0210042092949152
+85,20037,0.013504604808986187,
+86,20049,,0.007207389455288649
+86,20099,,0.00864070001989603
+86,20149,,0.011268459260463715
+86,20199,,0.007223770022392273
+86,20249,,0.01171328779309988
+86,20270,0.012282353825867176,
+87,20299,,0.03952428326010704
+87,20349,,0.013649888336658478
+87,20399,,0.009539869613945484
+87,20449,,0.010559165850281715
+87,20499,,0.020913757383823395
+87,20503,0.01216243114322424,
+88,20549,,0.00974198803305626
+88,20599,,0.008700666017830372
+88,20649,,0.0136112654581666
+88,20699,,0.00631662318482995
+88,20736,0.011909127235412598,
+89,20749,,0.00828890223056078
+89,20799,,0.008095502853393555
+89,20849,,0.009632572531700134
+89,20899,,0.009283551014959812
+89,20949,,0.009867608547210693
+89,20969,0.011236408725380898,
+90,20999,,0.010244566947221756
+90,21049,,0.014270391315221786
+90,21099,,0.008844885975122452
+90,21149,,0.010465869680047035
+90,21199,,0.020097170025110245
+90,21202,0.01119254156947136,
+91,21249,,0.008599667809903622
+91,21299,,0.008887047879397869
+91,21349,,0.00966575276106596
+91,21399,,0.0483819954097271
+91,21435,0.01116661075502634,
+92,21449,,0.008002564311027527
+92,21499,,0.0077667152509093285
+92,21549,,0.01422396395355463
+92,21599,,0.017090551555156708
+92,21649,,0.007748427335172892
+92,21668,0.010742527432739735,
+93,21699,,0.007330714259296656
+93,21749,,0.0067668333649635315
+93,21799,,0.01853134296834469
+93,21849,,0.010422656312584877
+93,21899,,0.004994779359549284
+93,21901,0.010509929619729519,
+94,21949,,0.005100528709590435
+94,21999,,0.011793495155870914
+94,22049,,0.007109381724148989
+94,22099,,0.009968251921236515
+94,22134,0.010136271826922894,
+95,22149,,0.007859839126467705
+95,22199,,0.014997732825577259
+95,22249,,0.013193503022193909
+95,22299,,0.007750023156404495
+95,22349,,0.01251092180609703
+95,22367,0.010636497288942337,
+96,22399,,0.00472773052752018
+96,22449,,0.010807463899254799
+96,22499,,0.008131821639835835
+96,22549,,0.0066901003010571
+96,22599,,0.009256426244974136
+96,22600,0.009883932769298553,
+97,22649,,0.012555822730064392
+97,22699,,0.009329312480986118
+97,22749,,0.005677599925547838
+97,22799,,0.00946832075715065
+97,22833,0.010213701985776424,
+98,22849,,0.013071042485535145
+98,22899,,0.01071584690362215
+98,22949,,0.013170363381505013
+98,22999,,0.010615585371851921
+98,23049,,0.025251155719161034
+98,23066,0.010050089098513126,
+99,23099,,0.021966515108942986
+99,23149,,0.007157896179705858
+99,23199,,0.009163573384284973
+99,23249,,0.010632473975419998
+99,23299,,0.008124726824462414
+99,23299,0.010232177563011646,
+100,23349,,0.018834497779607773
+100,23399,,0.007206527981907129
+100,23449,,0.009603036567568779
+100,23499,,0.008565641939640045
+100,23532,0.009783160872757435,
+101,23549,,0.007550579961389303
+101,23599,,0.011143269948661327
+101,23649,,0.00864416267722845
+101,23699,,0.006763931829482317
+101,23749,,0.011101005598902702
+101,23765,0.00949318427592516,
+102,23799,,0.008981946855783463
+102,23849,,0.004777817986905575
+102,23899,,0.01364970300346613
+102,23949,,0.011660045012831688
+102,23998,0.009460101835429668,
+103,23999,,0.008608500473201275
+103,24049,,0.009194415993988514
+103,24099,,0.00756433280184865
+103,24149,,0.006378332618623972
+103,24199,,0.009131166152656078
+103,24231,0.009468501433730125,
+104,24249,,0.007010235451161861
+104,24299,,0.007963284850120544
+104,24349,,0.05550863966345787
+104,24399,,0.006463226862251759
+104,24449,,0.010473793372511864
+104,24464,0.009219811297953129,
+105,24499,,0.007868734188377857
+105,24549,,0.007729876786470413
+105,24599,,0.006053104065358639
+105,24649,,0.006294563878327608
+105,24697,0.00937480665743351,
+106,24699,,0.008584842085838318
+106,24749,,0.012170390225946903
+106,24799,,0.006558562628924847
+106,24849,,0.006365340668708086
+106,24899,,0.009135046973824501
+106,24930,0.009154938161373138,
+107,24949,,0.013552608899772167
+107,24999,,0.008870783261954784
+107,25049,,0.005260922014713287
+107,25099,,0.009735574945807457
+107,25149,,0.015945330262184143
+107,25163,0.00921320728957653,
+108,25199,,0.017853258177638054
+108,25249,,0.012140416540205479
+108,25299,,0.004484710283577442
+108,25349,,0.008458084426820278
+108,25396,0.009133687242865562,
+109,25399,,0.0071701956912875175
+109,25449,,0.006333579774945974
+109,25499,,0.007580679841339588
+109,25549,,0.005078890360891819
+109,25599,,0.0065825823694467545
+109,25629,0.00919408816844225,
+110,25649,,0.008064880967140198
+110,25699,,0.00929039716720581
+110,25749,,0.009787668474018574
+110,25799,,0.007239412050694227
+110,25849,,0.007461087312549353
+110,25862,0.009005999192595482,
+111,25899,,0.008147509768605232
+111,25949,,0.004873862490057945
+111,25999,,0.007436844054609537
+111,26049,,0.008126542903482914
+111,26095,0.008990668691694736,
+112,26099,,0.008499368093907833
+112,26149,,0.007514809723943472
+112,26199,,0.005465700291097164
+112,26249,,0.007862327620387077
+112,26299,,0.005882487166672945
+112,26328,0.00888657197356224,
+113,26349,,0.0056334082037210464
+113,26399,,0.00853637233376503
+113,26449,,0.0047028036788105965
+113,26499,,0.007765656802803278
+113,26549,,0.007744417991489172
+113,26561,0.008866281248629093,
+114,26599,,0.007275634445250034
+114,26649,,0.007499241270124912
+114,26699,,0.012734834104776382
+114,26749,,0.007764542009681463
+114,26794,0.008873171173036098,
+115,26799,,0.005998617969453335
+115,26849,,0.006248430348932743
+115,26899,,0.004960114136338234
+115,26949,,0.0053023225627839565
+115,26999,,0.007784090470522642
+115,27027,0.008631293661892414,
+116,27049,,0.020947542041540146
+116,27099,,0.007255012169480324
+116,27149,,0.005884598009288311
+116,27199,,0.005914670415222645
+116,27249,,0.013698103837668896
+116,27260,0.00901118665933609,
+117,27299,,0.010400929488241673
+117,27349,,0.0060319602489471436
+117,27399,,0.017185894772410393
+117,27449,,0.006376294419169426
+117,27493,0.008654482662677765,
+118,27499,,0.006112037226557732
+118,27549,,0.009709403850138187
+118,27599,,0.011542784981429577
+118,27649,,0.006316347513347864
+118,27699,,0.007380709517747164
+118,27726,0.00860694982111454,
+119,27749,,0.0051675173453986645
+119,27799,,0.008363517001271248
+119,27849,,0.006931686773896217
+119,27899,,0.011661438271403313
+119,27949,,0.008507467806339264
+119,27959,0.00849993247538805,
+120,27999,,0.01423876266926527
+120,28049,,0.007003788836300373
+120,28099,,0.007078069262206554
+120,28149,,0.013359297066926956
+120,28192,0.00836015772074461,
+121,28199,,0.009029611013829708
+121,28249,,0.006018614862114191
+121,28299,,0.0072647156193852425
+121,28349,,0.005743059329688549
+121,28399,,0.0063291615806519985
+121,28425,0.008449426852166653,
+122,28449,,0.011916104704141617
+122,28499,,0.006538670975714922
+122,28549,,0.006099531892687082
+122,28599,,0.007182166446000338
+122,28649,,0.006912669632583857
+122,28658,0.008481699042022228,
+123,28699,,0.0071630701422691345
+123,28749,,0.005676476284861565
+123,28799,,0.008322194218635559
+123,28849,,0.006588925141841173
+123,28891,0.008385318331420422,
+124,28899,,0.006715129129588604
+124,28949,,0.0046298266388475895
+124,28999,,0.0045791976153850555
+124,29049,,0.005021781660616398
+124,29099,,0.015924064442515373
+124,29124,0.008418125100433826,
+125,29149,,0.008916601538658142
+125,29199,,0.006212690379470587
+125,29249,,0.005429331678897142
+125,29299,,0.007248596753925085
+125,29349,,0.007207055110484362
+125,29357,0.008279496803879738,
+126,29399,,0.006323528476059437
+126,29449,,0.006716982461512089
+126,29499,,0.005460174288600683
+126,29549,,0.004798295442014933
+126,29590,0.00802029762417078,
+127,29599,,0.010112179443240166
+127,29649,,0.0054520536214113235
+127,29699,,0.008057708851993084
+127,29749,,0.005594226997345686
+127,29799,,0.009124589152634144
+127,29823,0.00787987932562828,
+128,29849,,0.006573852617293596
+128,29899,,0.005986103322356939
+128,29949,,0.00629796739667654
+128,29999,,0.006925624329596758
+128,30049,,0.028203174471855164
+128,30056,0.008218895643949509,
+129,30099,,0.005303604528307915
+129,30149,,0.00823366828262806
+129,30199,,0.008076443336904049
+129,30249,,0.005666161887347698
+129,30289,0.008095948025584221,
+130,30299,,0.006363274063915014
+130,30349,,0.010425647720694542
+130,30399,,0.006245138123631477
+130,30449,,0.01085128914564848
+130,30499,,0.008620311506092548
+130,30522,0.008161315694451332,
+131,30549,,0.008366824127733707
+131,30599,,0.009563595987856388
+131,30649,,0.010114502161741257
+131,30699,,0.006071112584322691
+131,30749,,0.005452034529298544
+131,30755,0.007972240447998047,
+132,30799,,0.00525331636890769
+132,30849,,0.010969644412398338
+132,30899,,0.007480324245989323
+132,30949,,0.006384207401424646
+132,30988,0.008127872832119465,
+133,30999,,0.006873930338770151
+133,31049,,0.00931630190461874
+133,31099,,0.0074219233356416225
+133,31149,,0.006929187104105949
+133,31199,,0.007220882922410965
+133,31221,0.008026429452002048,
+134,31249,,0.00726038683205843
+134,31299,,0.016322379931807518
+134,31349,,0.006356596015393734
+134,31399,,0.005578037817031145
+134,31449,,0.008413984440267086
+134,31454,0.008534874767065048,
+135,31499,,0.013690531253814697
+135,31549,,0.006001586560159922
+135,31599,,0.00640503317117691
+135,31649,,0.011542258784174919
+135,31687,0.008443756029009819,
+136,31699,,0.0067335874773561954
+136,31749,,0.005280566867440939
+136,31799,,0.005492171738296747
+136,31849,,0.0069390395656228065
+136,31899,,0.008835828863084316
+136,31920,0.008214224129915237,
+137,31949,,0.01806776411831379
+137,31999,,0.005667026620358229
+137,32049,,0.0061721461825072765
+137,32099,,0.005701934453099966
+137,32149,,0.012365772388875484
+137,32153,0.008296037092804909,
+138,32199,,0.008139712736010551
+138,32249,,0.005744713358581066
+138,32299,,0.008589616045355797
+138,32349,,0.007126661483198404
+138,32386,0.008309477008879185,
+139,32399,,0.0070837922394275665
+139,32449,,0.004870944656431675
+139,32499,,0.007155123632401228
+139,32549,,0.006377063225954771
+139,32599,,0.006050293333828449
+139,32619,0.008268756791949272,
+140,32649,,0.008555719628930092
+140,32699,,0.008807959035038948
+140,32749,,0.008072678931057453
+140,32799,,0.008286567404866219
+140,32849,,0.013317829929292202
+140,32852,0.008409185335040092,

anvil_training/model.json

@@ -0,0 +1,14 @@
+{
+  "n_tasks": 5,
+  "ffn_hidden_dim": 512,
+  "batch_norm": false,
+  "dropout": 0.25,
+  "from_chemeleon": true,
+  "scheduler": "plateau",
+  "mpnn_lr": 0.0001,
+  "ffn_lr": 0.001,
+  "mpnn_weight_decay": 0.0,
+  "ffn_weight_decay": 0.0001,
+  "reduce_lr_factor": 0.5,
+  "reduce_lr_patience": 5
+}

anvil_training/model.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23a9b6fbb9c1db81555835eaf4ee78b485fb1a9c6d1deef689ab95998efe2c2b
+size 39069388

anvil_training/recipe_components/data.yaml

@@ -0,0 +1,15 @@
+anvil_dir: file:///home/ec2-user/permeability_models/multitask/all_endpoints_model
+cat_entry: null
+dropna: false
+input_col: OPENADMET_CANONICAL_SMILES
+resource: ../../ChEMBL35_Caco2_permeability_multitask_atob_btoa_logD_mppb_rescaled.parquet
+target_cols:
+- logD
+- caco2_atob_LogPapp
+- caco2_btoa_LogPapp
+- mppb_LogUnbound
+- hppb_LogUnbound
+test_resource: null
+train_resource: null
+type: intake
+val_resource: null

anvil_training/recipe_components/eval.yaml

@@ -0,0 +1,13 @@
+eval:
+- params: {}
+  type: RegressionMetrics
+- params:
+    axes_labels:
+    - True LogPapp
+    - Predicted LogPapp
+    n_repeats: 5
+    n_splits: 5
+    pXC50: true
+    random_state: 42
+    title: Multitask True vs Predicted LogPapp on test set
+  type: PytorchLightningRepeatedKFoldCrossValidation

anvil_training/recipe_components/metadata.yaml

@@ -0,0 +1,14 @@
+authors: Maria A. Castellanos
+biotargets:
+- Caco-2
+build_number: 0
+description: basic regression using a ChemProp multitask task model
+driver: pytorch
+email: maria.castellanos@omsf.io
+name: chemprop_pchembl
+tag: chemprop
+tags:
+- openadmet
+- example
+- chemprop
+version: v1

anvil_training/recipe_components/procedure.yaml

@@ -0,0 +1,47 @@
+ensemble: null
+feat:
+  params:
+    batch_size: 128
+    n_jobs: 4
+    normalize_targets: true
+  type: ChemPropFeaturizer
+model:
+  freeze_weights: null
+  param_path: null
+  params:
+    batch_norm: false
+    dropout: 0.25
+    ffn_hidden_dim: 512
+    ffn_hidden_num_layers: 3
+    ffn_lr: 1e-3
+    ffn_weight_decay: 1e-4
+    from_chemeleon: true
+    mpnn_lr: 1e-4
+    mpnn_weight_decay: 0
+    n_tasks: 5
+    reduce_lr_factor: 0.5
+    reduce_lr_patience: 5
+    scheduler: plateau
+  serial_path: null
+  type: ChemPropModel
+split:
+  params:
+    random_state: 42
+    test_size: 0.0
+    train_size: 1.0
+    val_size: 0.0
+  type: ShuffleSplitter
+train:
+  params:
+    accelerator: gpu
+    early_stopping: true
+    early_stopping_min_delta: 0.001
+    early_stopping_mode: min
+    early_stopping_patience: 20
+    gradient_clip_val: 0.5
+    max_epochs: 200
+    monitor_metric: val_loss
+    use_wandb: false
+    wandb_project: demos
+  type: LightningTrainer
+transform: null

meta.yaml

@@ -0,0 +1 @@
+# just for Hugging Face download tracking purposes

run_model_inference.sh

@@ -0,0 +1,16 @@
+#!/bin/bash
+# Script to run OpenADMET predictions in a standalone shell
+
+# Exit immediately if a command exits with a non-zero status
+set -e
+
+# --- Environment variables ---
+export TABPFN_TELEMETRY_OPTOUT=1
+export OADMET_NO_RICH_LOGGING=1
+
+openadmet predict \
+    --input-path expansion_data_inference.csv \
+    --input-col SMILES \
+    --model-dir anvil_training/ \
+    --output-csv predictions.csv \
+    --accelerator cpu