Add files using upload-large-folder tool

README.md

@@ -0,0 +1,80 @@
+---
+library_name: gridfm-graphkit
+pipeline_tag: tabular-regression
+license: apache-2.0
+tags:
+  - power-systems
+  - optimal-power-flow
+  - genco
+  - opfdata
+---
+
+# GENCO Base on OPFData (paper §5.2)
+
+Eighteen GENCO Base checkpoints used for Table 5 of [GENCO](https://arxiv.org/abs/2608.09921): six OPFData grids, three seeds (`42`, `3`, `17`).
+
+HH-MPNN numbers in the paper are taken from Arowolo et al.; they are not in this repo.
+
+## Files
+
+```
+{case14_ieee,case30_ieee,case57_ieee,case118_ieee,case500_goc,case2000_goc}/seed{42,3,17}/
+  best_model_state_dict.pt
+  normalizer_stats.pt
+  metrics.csv
+splits/
+  train.pt  val.pt  test.pt  shuffled_indices.pt
+mlflow/
+  train/<grid>/seed{42,3,17}/<run_id>/
+  eval/<run_id>/
+```
+
+`metrics.csv` is the eval test-set scalar dump for that checkpoint (`gridfm_graphkit evaluate --batch_size 512`). `splits/` are the HH-MPNN 90/5/5 index files (270k / 15k / 15k). Put them at `scripts/opfdata/splits/` when training or evaluating from [gridfm-graphkit `genco-paper-repro`](https://github.com/gridfm/gridfm-graphkit/tree/genco-paper-repro).
+
+The [training and eval MLflow runs](https://huggingface.co/gridfm/genco-opfdata-base/tree/main/mlflow) are under `mlflow/` (params, tags, metrics, test CSVs and plots). Weights stay in `<grid>/seed*/`, not inside the run dirs. Table 5 uses the **eval** runs, not the in-training test dump.
+
+## Install
+
+The paper results were obtained with [`genco-paper-repro`](https://github.com/gridfm/gridfm-graphkit/tree/genco-paper-repro). We only guarantee the same numbers on this branch, because `main` is under active development. Clone it and install from source:
+
+```bash
+git clone --branch genco-paper-repro https://github.com/gridfm/gridfm-graphkit.git
+cd gridfm-graphkit
+python -m venv venv && source venv/bin/activate
+pip install -e .
+TORCH_CUDA_VERSION=$(python -c "import torch; print(torch.__version__ + ('+cpu' if torch.version.cuda is None else ''))")
+pip install torch-scatter -f https://data.pyg.org/whl/torch-${TORCH_CUDA_VERSION}.html
+```
+
+## Data
+
+Converted OPFData parquet (already public). Graphkit loads `{data_path}/{network}/raw/*.parquet`.
+
+| Grid | Dataset |
+|---|---|
+| IEEE 14 | [gridfm/opfdata_case14_ieee](https://huggingface.co/datasets/gridfm/opfdata_case14_ieee) |
+| IEEE 30 | [gridfm/opfdata_case30_ieee](https://huggingface.co/datasets/gridfm/opfdata_case30_ieee) |
+| IEEE 57 | [gridfm/opfdata_case57_ieee](https://huggingface.co/datasets/gridfm/opfdata_case57_ieee) |
+| IEEE 118 | [gridfm/opfdata_case118_ieee](https://huggingface.co/datasets/gridfm/opfdata_case118_ieee) |
+| GOC 500 | [gridfm/opfdata_case500_goc](https://huggingface.co/datasets/gridfm/opfdata_case500_goc) |
+| GOC 2000 | [gridfm/opfdata_case2000_goc](https://huggingface.co/datasets/gridfm/opfdata_case2000_goc) |
+
+## Evaluate
+
+From the graphkit repo root:
+
+```bash
+mkdir -p data/case118_ieee/raw scripts/opfdata/splits
+hf download gridfm/opfdata_case118_ieee --repo-type dataset --local-dir data/case118_ieee/raw
+hf download gridfm/genco-opfdata-base --include "case118_ieee/seed42/**" --include "splits/**" --local-dir genco-opfdata-base
+cp genco-opfdata-base/splits/*.pt scripts/opfdata/splits/
+
+gridfm_graphkit evaluate \
+  --config scripts/opfdata/configs/HGNSQ_penalty_11_OPFData_case118_default.yaml \
+  --data_path data \
+  --model_path genco-opfdata-base/case118_ieee/seed42/best_model_state_dict.pt \
+  --normalizer_stats genco-opfdata-base/case118_ieee/seed42/normalizer_stats.pt \
+  --batch_size 512
+```
+
+Full commands: [GENCO `repro/OPF_on_opfdata_5.2`](https://github.com/albanpuech/GENCO/tree/main/repro/5.2).

case118_ieee/seed17/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.01797565445303917
+Avg. reactive res. (MVar),0.004483649041503668
+RMSE PG generators (MW),1.7834719134643902
+Mean optimality gap (%),0.1676466166973114
+Mean branch thermal violation from (MVA),0.004407372325658798
+Mean branch thermal violation to (MVA),0.00442260829731822
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.001608941936865449
+Mean Qg violation REF buses,0.0
+Mean Qg violation,0.001579146832227707

case118_ieee/seed3/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.01117081381380558
+Avg. reactive res. (MVar),0.002832532161846757
+RMSE PG generators (MW),1.7274079475222655
+Mean optimality gap (%),0.13894441723823547
+Mean branch thermal violation from (MVA),0.003641358111053705
+Mean branch thermal violation to (MVA),0.0036667746026068926
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.00241287425160408
+Mean Qg violation REF buses,0.0
+Mean Qg violation,0.0023681914899498224

case118_ieee/seed42/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.03642617166042328
+Avg. reactive res. (MVar),0.00811777263879776
+RMSE PG generators (MW),1.9721150175099342
+Mean optimality gap (%),0.18214885890483856
+Mean branch thermal violation from (MVA),0.005062632728368044
+Mean branch thermal violation to (MVA),0.00507196644321084
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.0021554548293352127
+Mean Qg violation REF buses,0.0
+Mean Qg violation,0.0021155388094484806

case14_ieee/seed17/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.000924755004234612
+Avg. reactive res. (MVar),0.0004367422079667449
+RMSE PG generators (MW),0.10347752499735767
+Mean optimality gap (%),0.06015359237790108
+Mean branch thermal violation from (MVA),0.0
+Mean branch thermal violation to (MVA),0.0
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.00977626908570528
+Mean Qg violation REF buses,0.005967199336737394
+Mean Qg violation,0.009014456532895565

case14_ieee/seed3/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.0007028199615888298
+Avg. reactive res. (MVar),0.0003320706309750676
+RMSE PG generators (MW),0.09211069601145432
+Mean optimality gap (%),0.05361694470047951
+Mean branch thermal violation from (MVA),0.0
+Mean branch thermal violation to (MVA),0.0
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.009303538128733635
+Mean Qg violation REF buses,0.007597099989652634
+Mean Qg violation,0.008962252177298069

case14_ieee/seed42/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.000853467034175992
+Avg. reactive res. (MVar),0.00038078916259109974
+RMSE PG generators (MW),0.09135909902153523
+Mean optimality gap (%),0.049257226288318634
+Mean branch thermal violation from (MVA),0.0
+Mean branch thermal violation to (MVA),0.0
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.009149911813437939
+Mean Qg violation REF buses,0.011247003450989723
+Mean Qg violation,0.009569330140948296

case2000_goc/seed17/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.06272722035646439
+Avg. reactive res. (MVar),0.023809390142560005
+RMSE PG generators (MW),1.5611490131149885
+Mean optimality gap (%),0.17134656012058258
+Mean branch thermal violation from (MVA),2.0847692212555557e-05
+Mean branch thermal violation to (MVA),2.3579481421620585e-05
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.0019401834579184651
+Mean Qg violation REF buses,0.0
+Mean Qg violation,0.001929283607751131

case2000_goc/seed3/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.05170949548482895
+Avg. reactive res. (MVar),0.019054226577281952
+RMSE PG generators (MW),1.5503898530257982
+Mean optimality gap (%),0.15517348051071167
+Mean branch thermal violation from (MVA),2.296143429703079e-05
+Mean branch thermal violation to (MVA),2.61342829617206e-05
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.0012824690202251077
+Mean Qg violation REF buses,0.0
+Mean Qg violation,0.001275264541618526

case2000_goc/seed42/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.06152397021651268
+Avg. reactive res. (MVar),0.02392873913049698
+RMSE PG generators (MW),1.6171995153303267
+Mean optimality gap (%),0.18053902685642242
+Mean branch thermal violation from (MVA),2.141216282325331e-05
+Mean branch thermal violation to (MVA),2.3745875296299346e-05
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.0017253225669264793
+Mean Qg violation REF buses,0.0
+Mean Qg violation,0.0017156299436464906

case30_ieee/seed17/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.001128296717070043
+Avg. reactive res. (MVar),0.0007495309109799564
+RMSE PG generators (MW),0.14233124273941128
+Mean optimality gap (%),0.031530290842056274
+Mean branch thermal violation from (MVA),0.001545486506074667
+Mean branch thermal violation to (MVA),0.0015151372645050287
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.0032189160119742155
+Mean Qg violation REF buses,0.001602351781912148
+Mean Qg violation,0.0029494885820895433

case30_ieee/seed3/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.0009840283310040832
+Avg. reactive res. (MVar),0.0007061019423417747
+RMSE PG generators (MW),0.16297330221802134
+Mean optimality gap (%),0.028041817247867584
+Mean branch thermal violation from (MVA),0.0012041671434417367
+Mean branch thermal violation to (MVA),0.0011926032602787018
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.0026464832480996847
+Mean Qg violation REF buses,0.0008685176144354045
+Mean Qg violation,0.0023501557298004627

case30_ieee/seed42/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.0010496790055185556
+Avg. reactive res. (MVar),0.0007321955636143684
+RMSE PG generators (MW),0.17869849239854416
+Mean optimality gap (%),0.030502844601869583
+Mean branch thermal violation from (MVA),0.001151254284195602
+Mean branch thermal violation to (MVA),0.0011356923496350646
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.0031471361871808767
+Mean Qg violation REF buses,0.0006270589074119925
+Mean Qg violation,0.0027271234430372715

case500_goc/seed17/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.060814470052719116
+Avg. reactive res. (MVar),0.024173736572265625
+RMSE PG generators (MW),2.59850160997938
+Mean optimality gap (%),0.29895922541618347
+Mean branch thermal violation from (MVA),0.0021828152239322662
+Mean branch thermal violation to (MVA),0.0021565707866102457
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.00033977837301790714
+Mean Qg violation REF buses,0.0
+Mean Qg violation,0.00033677142346277833

case500_goc/seed17/normalizer_stats.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e336f4e20127e9992e9a6640db85547f2b9aebfb5b5cd53925c7198cbe32abb
+size 2162

case500_goc/seed3/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.07679212838411331
+Avg. reactive res. (MVar),0.02644193544983864
+RMSE PG generators (MW),2.49663145097344
+Mean optimality gap (%),0.2589920461177826
+Mean branch thermal violation from (MVA),0.0018671727739274502
+Mean branch thermal violation to (MVA),0.001818525604903698
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.0004977736971341074
+Mean Qg violation REF buses,0.0004286865296307951
+Mean Qg violation,0.0004971624002791941

case500_goc/seed3/normalizer_stats.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e336f4e20127e9992e9a6640db85547f2b9aebfb5b5cd53925c7198cbe32abb
+size 2162

case500_goc/seed42/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.07300211489200592
+Avg. reactive res. (MVar),0.027255842462182045
+RMSE PG generators (MW),2.6076598783873863
+Mean optimality gap (%),0.33787015080451965
+Mean branch thermal violation from (MVA),0.0016855921130627394
+Mean branch thermal violation to (MVA),0.00165797327645123
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.0005428409785963595
+Mean Qg violation REF buses,0.0
+Mean Qg violation,0.000538037100341171

case500_goc/seed42/normalizer_stats.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e336f4e20127e9992e9a6640db85547f2b9aebfb5b5cd53925c7198cbe32abb
+size 2162

case57_ieee/seed17/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.038277022540569305
+Avg. reactive res. (MVar),0.023632513359189034
+RMSE PG generators (MW),3.058874202649445
+Mean optimality gap (%),0.09491429477930069
+Mean branch thermal violation from (MVA),2.6353896487307793e-07
+Mean branch thermal violation to (MVA),1.7296721921411518e-07
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.01232620608061552
+Mean Qg violation REF buses,0.002750933403149247
+Mean Qg violation,0.010958313010632992

case57_ieee/seed17/normalizer_stats.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50cfa761fbb0c8dcce05af711e5b8b09163c06f63d42cfd7234197e47e04da9e
+size 2162

case57_ieee/seed3/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.05140424892306328
+Avg. reactive res. (MVar),0.02881823293864727
+RMSE PG generators (MW),3.6052735380183774
+Mean optimality gap (%),0.11287030577659607
+Mean branch thermal violation from (MVA),4.5090177991369274e-07
+Mean branch thermal violation to (MVA),2.274615781061584e-06
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.008936013095080853
+Mean Qg violation REF buses,0.003405127441510558
+Mean Qg violation,0.008145886473357677

case57_ieee/seed3/normalizer_stats.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50cfa761fbb0c8dcce05af711e5b8b09163c06f63d42cfd7234197e47e04da9e
+size 2162

case57_ieee/seed42/metrics.csv

@@ -0,0 +1,11 @@
+Metric,Value
+Avg. active res. (MW),0.06472043693065643
+Avg. reactive res. (MVar),0.036860592663288116
+RMSE PG generators (MW),5.1295912111641355
+Mean optimality gap (%),0.13765496015548706
+Mean branch thermal violation from (MVA),7.882737804720819e-07
+Mean branch thermal violation to (MVA),9.720822617964586e-07
+Mean branch angle difference violation (radians),0.0
+Mean Qg violation PV buses,0.02058558166027069
+Mean Qg violation REF buses,0.0009447219781577587
+Mean Qg violation,0.017779747024178505

case57_ieee/seed42/normalizer_stats.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50cfa761fbb0c8dcce05af711e5b8b09163c06f63d42cfd7234197e47e04da9e
+size 2162

mlflow/README.md

@@ -0,0 +1,8 @@
+These are the paper §5.2 MLflow runs.
+
+- `train/<grid>/seed{42,3,17}/<run_id>/` — training runs that produced the checkpoints
+- `eval/<run_id>/` — `evaluate --batch_size 512` runs (`eval_3_seeds_paper_05_07`) that produced Table 5
+
+Saved `.pt` weights and `normalizer_stats.pt` live under `<grid>/seed*/`, not inside these run dirs.
+
+Train-run `artifacts/test/*_metrics.csv` is the in-training test dump. Table 5 uses the eval-run test metrics (copied to `<grid>/seed*/metrics.csv`).

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/meta.yaml

@@ -0,0 +1,14 @@
+artifact_uri: file:///dccstor/gridfm/march_opf_exp/results/840902984492823557/01ec28362fc8462aacd9e53df9614815/artifacts
+end_time: 1775470254717
+entry_point_name: ''
+experiment_id: '840902984492823557'
+lifecycle_stage: active
+run_id: 01ec28362fc8462aacd9e53df9614815
+run_name: run
+source_name: ''
+source_type: 4
+source_version: ''
+start_time: 1775074150814
+status: 3
+tags: []
+user_id: zrltbr

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/callbacks/optuna_early_prune

@@ -0,0 +1 @@
+False

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/callbacks/patience

@@ -0,0 +1 @@
+100

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/callbacks/tol

@@ -0,0 +1 @@
+0

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/data/baseMVA

@@ -0,0 +1 @@
+100

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/data/mask_value

@@ -0,0 +1 @@
+0.0

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/data/networks

@@ -0,0 +1 @@
+['case500_goc']

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/data/normalization

@@ -0,0 +1 @@
+HeteroDataMVANormalizer

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/data/scenarios

@@ -0,0 +1 @@
+[300000]

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/data/split_from_existing_files

@@ -0,0 +1 @@
+/dccstor/gridfm/march_opf_exp/opfdata_olay_splits/

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/data/test_ratio

@@ -0,0 +1 @@
+0.05

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/data/val_ratio

@@ -0,0 +1 @@
+0.05

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/data/workers

@@ -0,0 +1 @@
+16

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/optimizer/learning_rate

@@ -0,0 +1 @@
+0.0005

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/optimizer/optimizer_params/betas

@@ -0,0 +1 @@
+[0.9, 0.999]

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/optimizer/scheduler_params/factor

@@ -0,0 +1 @@
+0.7

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/optimizer/scheduler_params/mode

@@ -0,0 +1 @@
+min

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/optimizer/scheduler_params/patience

@@ -0,0 +1 @@
+5

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/optimizer/scheduler_type

@@ -0,0 +1 @@
+ReduceLROnPlateau

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/optimizer/type

@@ -0,0 +1 @@
+AdamW

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/seed

@@ -0,0 +1 @@
+17

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/task/task_name

@@ -0,0 +1 @@
+OptimalPowerFlow

mlflow/train/case500_goc/seed17/01ec28362fc8462aacd9e53df9614815/params/args/verbose

@@ -0,0 +1 @@
+True

mlflow/train/case500_goc/seed17/meta.yaml

@@ -0,0 +1,7 @@
+artifact_location: file:///dccstor/gridfm/march_opf_exp/results/840902984492823557
+creation_time: 1775074149273
+experiment_id: '840902984492823557'
+last_update_time: 1775074149273
+lifecycle_stage: active
+name: HGNSQ_penalty_11_OPFData_case500_17
+workspace: default