miangangzhen

I think you need to modify the evaluation function. We should chose the model that work the best on evaluation data

Very simple, just like run it on PC or server

Please try use my code, that read csv file with text and summary. Replace that with the Batcher. ```python def example_generator(data_path, single_pass): while True: filelist = tf.gfile.Glob(data_path) assert filelist, ('Error:...

Just add some code, like tf.logging.info(train_step) ```python # get the summaries and iteration number so we can write summaries to tensorboard summaries = results['summaries'] # we will write these summaries...

Solution: add these code in batcher.py ======== before ============== ```python def fill_example_queue(self): input_gen = self.text_generator(data.example_generator(self._data_path, self._single_pass)) while True: try: (article, abstract) = next(input_gen) ``` ========after================ ```python def fill_example_queue(self): input_gen =...

I encounter same problem, and i upgrade mcp package: pip3 install mcp -U