refactor: 检查点加载重构，路径替代对象传递

- model: nn.Module -> model_fn 工厂函数，spawn 边界只传字符串
- Trainer.train(resume_dir=path) — Checkpoint 不再通过 pickle 传递
- TrainContextBuilder.with_resume_dir(path) — 自动检测 meta.json 分流 resume/from-scratch
- CheckpointCallback: 拆分 state_dict 收集（全 rank）与磁盘写入（rank-0），修复 FSDP 死锁
- serialization: load_torch 支持 broadcast，消除 _load_extra/_load_torch_broadcast
- optimizer/scheduler 恢复逻辑内联到 build()，在 executor.prepare() 之后执行
- pyproject.toml: ruff exclude build/ 避免 CI 扫描构建产物

tests/trainer/test_callbacks.py

@@ -106,7 +106,7 @@ def test_gradient_checkpointing_trainer_integration(base_test_env, random_datase
         )
 
     train_config = TrainConfig(
-        model=base_test_env["model"],
+        model_fn=lambda: base_test_env["model"],
         strategy="seq",
         dataset=random_dataset,
         optimizer_fn=optimizer_fn,
@@ -140,7 +140,7 @@ def test_callback_integration(base_test_env, random_dataset):
         )
 
     train_config = TrainConfig(
-        model=base_test_env["model"],
+        model_fn=lambda: base_test_env["model"],
         strategy="seq",
         dataset=random_dataset,
         optimizer_fn=optimizer_fn,