def eval()

in multiple_futures_prediction/train_ngsim.py [0:0]

• 42 lines of code
• 12 McCabe index (conditional complexity)

def eval(metric: str, net: torch.nn.Module, params: AttrDict, data_loader: DataLoader, bStepByStep: bool, 
        use_forcing: int, y_mean: np.ndarray, num_batches: int, dataset_name: str) -> torch.Tensor:
  """Evaluation function for validation and test data.  
  
  Given a MFP network, data loader, evaulate either NLL or RMSE error.
  """  
  print('eval ', dataset_name)
  num = params.fut_len_orig_hz//params.subsampling
  lossVals = torch.zeros(num)
  counts = torch.zeros(num)

  for i, data in enumerate(data_loader):
    if i >= num_batches:
      break      
    hist, nbrs, mask, fut, mask, context, nbrs_info = data
    if params.use_cuda:
      hist = hist.cuda()
      nbrs = nbrs.cuda()
      mask = mask.cuda()            
      fut = fut.cuda()
      mask = mask.cuda()
      if context is not None:
        context = context.cuda()

    if metric == 'nll':      
      fut_preds, modes_pred = net.forward_mfp(hist, nbrs, mask, context, nbrs_info, fut, bStepByStep, use_forcing=use_forcing)
      if params.modes == 1:
        if params.remove_y_mean:
          fut_preds[0][:,:,:2] += y_mean.unsqueeze(1).to(fut.device)
        l, c = nll_loss_test(fut_preds[0], fut, mask)
      else:
        l, c = nll_loss_test_multimodes(fut_preds, fut, mask, modes_pred, y_mean.to(fut.device) )                        
    else: # RMSE error
      assert params.modes == 1
      fut_preds, modes_pred = net.forward_mfp(hist, nbrs, mask, context, nbrs_info, fut, bStepByStep, use_forcing=use_forcing)
      if params.modes == 1:
        if params.remove_y_mean:
          fut_preds[0][:,:,:2] += y_mean.unsqueeze(1).to(fut.device)
        l, c = mse_loss_test(fut_preds[0], fut, mask)

    lossVals  += l.detach().cpu()
    counts    += c.detach().cpu()

  if metric == 'nll':
    err = lossVals / counts
    print(lossVals / counts)
  else:
    err = torch.pow(lossVals / counts,0.5)*0.3048
    print( err ) # Calculate RMSE and convert from feet to meters
  return err