报一个小bug，用num_fakes做切片时没有取整。

[131250208]

报错信息：

TypeError                                 Traceback (most recent call last)
<ipython-input-7-039e026fb8a7> in <module>
     30 # step 8: predict
     31 print('predicting..')
---> 32 results = trainer.predict(print_steps=print_steps)

~/anaconda3/envs/paddle/lib/python3.6/site-packages/paddlepalm/trainer.py in predict(self, output_dir, print_steps)
    645         cur_predict_step = 0
    646         for feed in iterator:
--> 647             rt_outputs = self.predict_one_batch(feed)
    648             # rt_outputs = {k[len(self.name+'.'):]: v for k,v in rt_outputs.items() if k.startswith(self.name+'.')}
    649             self._pred_head.batch_postprocess(rt_outputs)

~/anaconda3/envs/paddle/lib/python3.6/site-packages/paddlepalm/trainer.py in predict_one_batch(self, batch)
    752             num_fakes = decode_fake(len(rt_outputs[0]), mask, self._predict_batch_size)
    753             if num_fakes:
--> 754                 rt_outputs = [i[:-num_fakes] for i in rt_outputs]
    755         else:
    756             feed = self._pred_feed_batch_process_fn(batch)

~/anaconda3/envs/paddle/lib/python3.6/site-packages/paddlepalm/trainer.py in <listcomp>(.0)
    752             num_fakes = decode_fake(len(rt_outputs[0]), mask, self._predict_batch_size)
    753             if num_fakes:
--> 754                 rt_outputs = [i[:-num_fakes] for i in rt_outputs]
    755         else:
    756             feed = self._pred_feed_batch_process_fn(batch)

TypeError: slice indices must be integers or None or have an __index__ method

Debug:

ipdb> num_fakes
6.0

[xixiaoyao]

非常感谢，bug已修复，可以update后再试一下