Update main-training.py

main-training.py

@@ -390,7 +390,7 @@ class CascadeConfig:
     beta2:        float = 0.95
     grad_clip:    float = 1.0
     warmup_steps: int   = 2000
-    total_steps:  int   = 80000
+    total_steps:  int   = 750000
 
     micro_batch: int = 4
     grad_accum:  int = 8
@@ -1661,15 +1661,15 @@ def train():
 
     # ── checkpoint resume (uncomment to resume within session 3) ─────────────
     start_step, start_epoch, best_val = 0, 0, float("inf")
-    # latest = find_latest_checkpoint()
-    # if latest:
-    #     try:
-    #         start_step, start_epoch, best_val = load_ckpt(
-    #             latest, model, muon_opt, adamw_opt, device, cfg
-    #         )
-    #     except Exception as e:
-    #         log.error(f"[ckpt] resume failed ({e}) — fresh start")
-    #         start_step = start_epoch = 0
+    latest = find_latest_checkpoint()
+    if latest:
+        try:
+            start_step, start_epoch, best_val = load_ckpt(
+                latest, model, muon_opt, adamw_opt, device, cfg
+            )
+        except Exception as e:
+            log.error(f"[ckpt] resume failed ({e}) — fresh start")
+            start_step = start_epoch = 0
 
     model.train()
     model.set_step(start_step)