网络训练至某个epoch，参数 问题

start_epoch = params.start_epoch
  stop_epoch = params.stop_epoch
  if params.resume != '':
    resume_file = get_resume_file('%s/checkpoints/%s'%(params.save_dir, params.resume), params.resume_epoch)  #get_resume_file函数得到epoch.tar文件
    if resume_file is not None:
      tmp = torch.load(resume_file)
      start_epoch = tmp['epoch']+1
      model.load_state_dict(tmp['state'])
      print('  resume the training with at {} epoch (model file {})'.format(start_epoch, params.resume))

对tar文件进行加载，并且选取其中需要的字典权重

tmp = torch.load(modelfile)   # load parameter file:400.tar
  try:
    state = tmp['state']
  except KeyError:
    state = tmp['model_state']
  except:
    raise
  state_keys = list(state.keys())  #列举字典中的key
  for i, key in enumerate(state_keys):
    if "feature." in key and not 'gamma' in key and not 'beta' in key:
      newkey = key.replace("feature.","")
      state[newkey] = state.pop(key)  #删除该key并返回对应的值，不影响上面的训练
    else:
      state.pop(key)

  model.load_state_dict(state)