PaddleOCR 加载预训练模型老出错

# 训练 mv3_db 模型，并将训练日志保存为 tain_det.log
%cd PaddleOCR/
!python3 tools/train.py -c configs/det/det_r50_vd_db.yml \
     -o Global.pretrain_weights=./pretrain_models/ResNet50_vd_ssld_pretrained \
     2>&1 | tee train_det.log

报错信息：

W1223 01:25:11.035835  4034 device_context.cc:447] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 10.1, Runtime API Version: 10.1
W1223 01:25:11.040609  4034 device_context.cc:465] device: 0, cuDNN Version: 7.6.
Traceback (most recent call last):
  File "tools/train.py", line 148, in 
    main(config, device, logger, vdl_writer)
  File "tools/train.py", line 100, in main
    pre_best_model_dict = load_model(config, model, optimizer)
  File "/home/aistudio/PaddleOCR/ppocr/utils/save_load.py", line 98, in load_model
    load_pretrained_params(model, pretrained_model)
  File "/home/aistudio/PaddleOCR/ppocr/utils/save_load.py", line 109, in load_pretrained_params
    "The {}.pdparams does not exists!".format(path)
AssertionError: The ./pretrain_models/ResNet50_vd_ssld_pretrained.pdparams does not exists!

就是说默认预训练模型都是*.pdparams 格式，对于散列模型，怎么读取？