diff options
Diffstat (limited to 'protocol/examples/temporal_diagnostic_evolution.py')
| -rw-r--r-- | protocol/examples/temporal_diagnostic_evolution.py | 34 |
1 files changed, 25 insertions, 9 deletions
diff --git a/protocol/examples/temporal_diagnostic_evolution.py b/protocol/examples/temporal_diagnostic_evolution.py index 6a2c042..35cf720 100644 --- a/protocol/examples/temporal_diagnostic_evolution.py +++ b/protocol/examples/temporal_diagnostic_evolution.py @@ -39,10 +39,16 @@ from protocol.report import DiagnosticThresholds # noqa: E402 THRESHOLDS = DiagnosticThresholds() +def max_per_block_growth(h): + if len(h) < 2: + return 1.0 + return max(h[i + 1] / max(h[i], 1e-30) for i in range(len(h) - 1)) + + def diagnose_entry(entry): h = entry["hidden_norms"] g = entry["bp_grad_norms_per_sample_med"] - h_exploded = (max(h) / max(h[0], 1e-30)) > THRESHOLDS.h_norm_explosion_ratio + h_exploded = max_per_block_growth(h) > THRESHOLDS.h_norm_explosion_ratio g_at_floor = g[-1] < THRESHOLDS.g_norm_floor return h_exploded, g_at_floor @@ -58,17 +64,27 @@ def main(): import argparse p = argparse.ArgumentParser() p.add_argument("--seed", type=int, default=42) + p.add_argument("--arch", type=str, default="resmlp", choices=["resmlp", "vit"]) args = p.parse_args() - snapshot_path = os.path.join( - REPO_ROOT, f"results/snapshot_evolution_v2/snapshot_evolution_s{args.seed}.json" - ) + if args.arch == "resmlp": + snapshot_path = os.path.join( + REPO_ROOT, f"results/snapshot_evolution_v2/snapshot_evolution_s{args.seed}.json" + ) + h_key = "hidden_norms" + g_key = "bp_grad_norms_per_sample_med" + else: + snapshot_path = os.path.join( + REPO_ROOT, f"results/snapshot_vit_v1/snapshot_vit_s{args.seed}.json" + ) + h_key = "hidden_norms_cls" + g_key = "bp_grad_per_sample_l2_med" if not os.path.exists(snapshot_path): print(f"snapshot not found: {snapshot_path}") return with open(snapshot_path) as f: d = json.load(f) - bp_log = d["bp_log"] - dfa_log = d["dfa_log"] + bp_log = [{**e, "hidden_norms": e[h_key], "bp_grad_norms_per_sample_med": e[g_key]} for e in d["bp_log"]] + dfa_log = [{**e, "hidden_norms": e[h_key], "bp_grad_norms_per_sample_med": e[g_key]} for e in d["dfa_log"]] print("=" * 88) print("TEMPORAL DIAGNOSTIC EVOLUTION (4-block d=256 ResMLP, CIFAR-10, seed 42)") @@ -87,7 +103,7 @@ def main(): for entry in dfa_log: h = entry["hidden_norms"] g = entry["bp_grad_norms_per_sample_med"] - h_exp = (max(h) / max(h[0], 1e-30)) > THRESHOLDS.h_norm_explosion_ratio + h_exp = max_per_block_growth(h) > THRESHOLDS.h_norm_explosion_ratio g_floor = g[-1] < THRESHOLDS.g_norm_floor flag_a = "FIRE" if h_exp else "ok" flag_b = "FIRE" if g_floor else "ok" @@ -125,7 +141,7 @@ def main(): for entry in bp_log: h = entry["hidden_norms"] g = entry["bp_grad_norms_per_sample_med"] - h_exp = (max(h) / max(h[0], 1e-30)) > THRESHOLDS.h_norm_explosion_ratio + h_exp = max_per_block_growth(h) > THRESHOLDS.h_norm_explosion_ratio g_floor = g[-1] < THRESHOLDS.g_norm_floor if h_exp or g_floor: bp_fired = True @@ -145,7 +161,7 @@ def main(): { "epoch": e["epoch"], "acc": e["acc_eval"], - "h_max_to_h0_ratio": (max(e["hidden_norms"]) / max(e["hidden_norms"][0], 1e-30)), + "max_per_block_growth": max_per_block_growth(e["hidden_norms"]), "g_L": e["bp_grad_norms_per_sample_med"][-1], "gamma": e.get("gamma_dfa"), } |