1 files changed, 101 insertions, 0 deletions
diff --git a/experiments/run_dfagnn_depth.py b/experiments/run_dfagnn_depth.py
new file mode 100644
index 0000000..ed6e6c3
--- /dev/null
+++ b/experiments/run_dfagnn_depth.py
@@ -0,0 +1,101 @@
+#!/usr/bin/env python3
+"""H7: DFA-GNN depth sweep for Figure 4(a)-style plot.
+
+Runs DFA-GNN at L ∈ {4, 8, 10, 12, 16, 20} × {Cora, CiteSeer, PubMed, DBLP} × 20 seeds.
+L=6 data already exists from prior experiments; L=2/3 skipped (CiteSeer L=2 GRAFT soft spot).
+
+Combined with existing BP and GRAFT depth data, produces 3-method depth curves for Figure 4(a).
+"""
+
+import torch
+import numpy as np
+import json
+import os
+from src.data import load_dataset
+from src.trainers import DFAGNNTrainer
+from run_dblp_depth import load_dblp
+
+device = 'cuda:0'
+SEEDS = list(range(20))
+EPOCHS = 200
+DEPTHS = [4, 8, 10, 12, 16, 20]
+OUT_DIR = 'results/dfagnn_depth_20seeds'
+
+dfagnn_extra = dict(diffusion_alpha=0.5, diffusion_iters=10, max_topo_power=3)
+
+
+def train_one(data, L, seed):
+    torch.manual_seed(seed); np.random.seed(seed); torch.cuda.manual_seed_all(seed)
+    t = DFAGNNTrainer(data=data, hidden_dim=64, lr=0.01, weight_decay=5e-4,
+                      num_layers=L, residual_alpha=0.0, backbone='gcn', **dfagnn_extra)
+    bv, bt = 0, 0
+    for ep in range(EPOCHS):
+        t.train_step()
+        if ep % 5 == 0:
+            v = t.evaluate('val_mask')
+            te = t.evaluate('test_mask')
+            if v > bv: bv, bt = v, te
+    del t; torch.cuda.empty_cache()
+    return bt
+
+
+def main():
+    os.makedirs(OUT_DIR, exist_ok=True)
+    per_seed_file = os.path.join(OUT_DIR, 'per_seed_data.json')
+    if os.path.exists(per_seed_file):
+        with open(per_seed_file) as f:
+            per_seed_data = json.load(f)
+    else:
+        per_seed_data = {}
+
+    datasets_cfg = {
+        'Cora': lambda: load_dataset('Cora', device=device),
+        'CiteSeer': lambda: load_dataset('CiteSeer', device=device),
+        'PubMed': lambda: load_dataset('PubMed', device=device),
+        'DBLP': lambda: load_dblp(),
+    }
+
+    for ds_name, loader in datasets_cfg.items():
+        data = loader()
+        for L in DEPTHS:
+            key = f"{ds_name}_L{L}_DFA-GNN"
+            if key not in per_seed_data:
+                per_seed_data[key] = {}
+
+            print(f"\n=== {key} (20 seeds) ===", flush=True)
+            for seed in SEEDS:
+                sk = str(seed)
+                if sk in per_seed_data[key]:
+                    print(f"  seed {seed}: cached ({per_seed_data[key][sk]*100:.1f}%)", flush=True)
+                    continue
+                try:
+                    acc = train_one(data, L, seed)
+                    per_seed_data[key][sk] = acc
+                    print(f"  seed {seed}: {acc*100:.1f}%", flush=True)
+                except Exception as e:
+                    print(f"  seed {seed}: FAILED - {e}", flush=True)
+                    per_seed_data[key][sk] = 0.0
+
+                with open(per_seed_file, 'w') as f:
+                    json.dump(per_seed_data, f, indent=2)
+        del data; torch.cuda.empty_cache()
+
+    # Summary
+    print(f"\n{'=' * 70}\nDFA-GNN depth sweep summary (20 seeds)\n{'=' * 70}")
+    results = {}
+    for ds in datasets_cfg:
+        print(f"\n{ds}:")
+        for L in DEPTHS:
+            key = f"{ds}_L{L}_DFA-GNN"
+            vals = np.array([per_seed_data[key][str(s)] for s in SEEDS]) * 100
+            results[key] = {'mean': float(vals.mean()), 'std': float(vals.std()),
+                             'per_seed': vals.tolist()}
+            print(f"  L={L:2d}  DFA-GNN {vals.mean():5.1f} ± {vals.std():4.1f}")
+
+    with open(os.path.join(OUT_DIR, 'results.json'), 'w') as f:
+        json.dump(results, f, indent=2)
+    print(f"\nSaved to {OUT_DIR}/results.json")
+
+
+if __name__ == '__main__':
+    main()