update fastNLP/core/tester.py.

开发者好，我在在0.7.0使用过程中发现了一些自己写的网络，在框架中是没有找到预测函数的，只单独返回预测结果。因此在tester分支中加入了一个只返回预测结果的函数，flp_topredict，并考虑到了不同divice转换。

fastNLP/core/tester.py

@@ -248,3 +248,52 @@ class Tester(object):
             _str += ", ".join([str(key) + "=" + str(value) for key, value in metric_result.items()])
             _str += '\n'
         return _str[:-1]
+
+    def flp_topredict(self):
+        r"""开始进行预测，并返回预测结果。
+
+        :return 本次的预测结果，为一个字典，其中只有{predict}一个key，而key的值类型为tensor。
+        """
+        # turn on the testing mode; clean up the history
+        self._model_device = _get_model_device(self._model)
+        network = self._model
+        self._mode(network, is_test=True)
+        data_iterator = self.data_iterator
+        eval_results = []
+        try:
+            with torch.no_grad():
+                if not self.use_tqdm:
+                    from .utils import _pseudo_tqdm as inner_tqdm
+                else:
+                    inner_tqdm = tqdm
+                with inner_tqdm(total=len(data_iterator), leave=False, dynamic_ncols=True) as pbar:
+                    pbar.set_description_str(desc="Pred")
+
+                    start_time = time.time()
+
+                    for batch_x, batch_y in data_iterator:
+                        _move_dict_value_to_device(batch_x, batch_y, device=self._model_device,
+                                                   non_blocking=self.pin_memory)
+                        with self.auto_cast():
+                            pred_dict = self._data_forward(self._predict_func, batch_x)
+
+                        eval_results.extend(pred_dict['predict'].detach().cpu().numpy())
+
+                        if self.use_tqdm:
+                            pbar.update()
+
+                    pbar.close()
+                    end_time = time.time()
+                    test_str = f'Predict data in {round(end_time - start_time, 2)} seconds!'
+                    if self.verbose >= 0:
+                        self.logger.info(test_str)
+        except _CheckError as e:
+            prev_func_signature = _get_func_signature(self._predict_func)
+            _check_loss_evaluate(prev_func_signature=prev_func_signature, func_signature=e.func_signature,
+                                 check_res=e.check_res, pred_dict=pred_dict, target_dict=batch_y,
+                                 dataset=self.data, check_level=0)
+        finally:
+            self._mode(network, is_test=False)
+            print(f'预测完成')
+
+        return eval_results