implememt cli

experiments/config.yaml

@@ -0,0 +1,48 @@
+# lightning.pytorch==2.5.5
+seed_everything: true
+trainer:
+  accelerator: gpu
+  strategy: auto
+  devices: 1
+  num_nodes: 1
+  precision: null
+  logger: null
+  callbacks:
+    - class_path: lightning.pytorch.callbacks.ModelCheckpoint
+      init_args:
+        monitor: val/loss
+        mode: min
+        save_top_k: 1
+        filename: best-checkpoint
+  max_epochs: 50
+  min_epochs: null
+  max_steps: -1
+  min_steps: null
+  overfit_batches: 0.0
+  log_every_n_steps: null
+  inference_mode: true
+  default_root_dir: null
+model: 
+  class_path: ThermalSolver.module.GraphSolver
+  init_args:
+    model_class_path: ThermalSolver.model.local_gno.GatingGNO
+    model_init_args:
+      x_ch_node: 1 
+      f_ch_node: 1 
+      hidden: 16 
+      layers: 2
+      edge_ch: 3 
+      out_ch: 1
+    unrolling_steps: 10
+data: 
+  class_path: ThermalSolver.data_module.GraphDataModule
+  init_args:
+    hf_repo: "SISSAmathLab/thermal-conduction"
+    split_name: "2000"
+    batch_size: 6
+    train_size: 0.8
+    test_size: 0.1
+    test_size: 0.1
+optimizer: null
+lr_scheduler: null
+ckpt_path: null

run.py

@@ -1,38 +1,12 @@
 import torch
-from lightning import Trainer
+from lightning.pytorch.cli import LightningCLI
-from ThermalSolver.module import GraphSolver
+
-from ThermalSolver.data_module import GraphDataModule
+torch.set_float32_matmul_precision("medium")
-from ThermalSolver.model.local_gno import GatingGNO
 
 
 def main():
-    trainer = Trainer(
+    LightningCLI(subclass_mode_data=True, subclass_mode_model=True)
-        max_epochs=50, accelerator="cuda", devices=1, accumulate_grad_batches=3
-    )
-    data_module = GraphDataModule(
-        hf_repo="SISSAmathLab/thermal-conduction",
-        split_name="2000",
-        train_size=0.8,
-        val_size=0.1,
-        test_size=0.1,
-        batch_size=10,
-    )
-    data_module.prepare_data()
-    data_module.setup("fit")
-    model = GatingGNO(
-        x_ch_node=1, f_ch_node=1, hidden=16, layers=2, edge_ch=3, out_ch=1
-    )
-    solver = GraphSolver(model, unrolling_steps=64)
-
-    trainer.fit(
-        solver,
-        train_dataloaders=data_module.train_dataloader(),
-        val_dataloaders=data_module.val_dataloader(),
-    )
-    data_module.setup("test")
-    trainer.test(solver, dataloaders=data_module.test_dataloader())
 
 
 if __name__ == "__main__":
-    torch.set_float32_matmul_precision("medium")
     main()