neural-network - 将张量输出保存到 csv 文件或至少一个数组

Question

batch_size = 48
train_data = TensorDataset(train_inputs, train_masks, train_labels)
#Training
train_loss_set = []
iterate = 4
for _ in trange(iterate, desc="Iterate"):
  model.train()

  train_loss = 0 
  nu_train_examples, nu_train_steps = 0, 0

  for step, batch in enumerate(train_dataloader):
    batch = tuple(t.to(device) for t in batch)
    batch_input_ids, batch_input_mask, batch_labels = batch
    optimizer.zero_grad()
    output = model(batch_input_ids, attention_mask=batch_input_mask)
    logits = output[0]

    logittt = logits[0]
    print(logittt[0],"\n")

和输出的一部分：

tensor(-0.4331, device='cuda:0', grad_fn=<SelectBackward>) 

tensor(-0.4924, device='cuda:0', grad_fn=<SelectBackward>) 

tensor(-0.4356, device='cuda:0', grad_fn=<SelectBackward>) 

tensor(-0.8334, device='cuda:0', grad_fn=<SelectBackward>) 

tensor(-0.6026, device='cuda:0', grad_fn=<SelectBackward>) 

tensor(-0.3859, device='cuda:0', grad_fn=<SelectBackward>)

这是我的分类代码的一部分。logittt 拥有一个 11x 数组。我只想要每个号码分别。例如 -0.4331 在第一个输出行中，以便添加到数组中。