fix bug for core dump in http, use libboost as parse (#2509)

zhaomingwork

2025-05-08 ae013cf597db1c523c9fac21b7e83db62304ae2d

 funasr/bin/train_ds.py

@@ -184,8 +184,10 @@
            )
            trainer.start_step = 0

            with torch.cuda.device(kwargs["device"]):
                torch.cuda.empty_cache()
            device = next(model.parameters()).device
            if device.type == "cuda":
                with torch.cuda.device(device):
                    torch.cuda.empty_cache()

            time_escaped = (time.perf_counter() - time_slice_i) / 3600.0
            logging.info(