| |
| """fastloop_merge_adapter.py <base-dir> <adapter-dir> <merged-dir> — re-create the |
| BF16 merged checkpoint from a kept adapter (the ladder runners delete merged/ |
| after eval; when an eval fails for plumbing reasons this rebuilds it in ~5 min |
| instead of a 2 h retune). Same merge law as fastloop_smoke_train.py M gate.""" |
| import os, shutil, sys, torch |
| from peft import PeftModel |
| from transformers import AutoConfig, AutoModelForCausalLM, AutoModelForImageTextToText, AutoTokenizer |
| base, adapter, merged_dir = sys.argv[1:4] |
| cfg = AutoConfig.from_pretrained(base) |
| cls = AutoModelForImageTextToText if hasattr(cfg, "text_config") else AutoModelForCausalLM |
| m = cls.from_pretrained(base, device_map="auto", dtype=torch.bfloat16) |
| m = PeftModel.from_pretrained(m, adapter).merge_and_unload() |
| if os.path.isdir(merged_dir): shutil.rmtree(merged_dir) |
| m.save_pretrained(merged_dir, safe_serialization=True) |
| AutoTokenizer.from_pretrained(base).save_pretrained(merged_dir) |
| for f in os.listdir(base): |
| if f.endswith(".index.json"): continue |
| if f.endswith((".jinja", ".json", ".py", ".txt", ".model")) and not os.path.exists(os.path.join(merged_dir, f)): |
| shutil.copy(os.path.join(base, f), merged_dir) |
| st = [f for f in os.listdir(merged_dir) if f.endswith(".safetensors")] |
| print(f"MERGE PASS: {len(st)} shards -> {merged_dir}" if st else "MERGE FAIL") |
| sys.exit(0 if st else 1) |
|
|
