item 24 pre-registered: d=1 retry with loop-only band-LoRA r=16 (4.8M params, whole band, k=0 bit-exact); trainer/eval gain --bandlora; smoke-tested train+eval

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

scripts/eval_carry_cot.py

@@ -36,11 +36,25 @@ def main():
     ap.add_argument("--batch", type=int, default=8)
     ap.add_argument("--feedforward", action="store_true")
     ap.add_argument("--max-new", type=int, default=160)
+    ap.add_argument("--bandlora", default=None, metavar="LORA_PT",
+                    help="load a lora_*_e*.pt loop-only band-LoRA checkpoint")
     args = ap.parse_args()
 
     model, tok = load_model(dtype=torch.bfloat16)
     tok.padding_side = "left"
     looper = BandLooper(model)
+    if args.bandlora:
+        from lora_band import inject_band_lora
+        ck = torch.load(args.bandlora, map_location="cuda")
+        scales = {l: 1.0 for l in ck["band"]}
+        ps = inject_band_lora(looper.tm, ck["band"][0], scales,
+                              rank=ck["rank"])
+        assert len(ps) == len(ck["tensors"]), (len(ps), len(ck["tensors"]))
+        for pr, t in zip(ps, ck["tensors"]):
+            pr.data = t.cuda()
+        print(f"band-lora loaded: {args.bandlora} "
+              f"(r={ck['rank']}, layers {ck['band'][0]}-{ck['band'][-1]})",
+              flush=True)
     adapter = MergeAdapter(
         d=model.config.get_text_config().hidden_size).cuda()
     adapter.load_state_dict(torch.load(args.adapter, map_location="cuda"))