tensorflow/models · error · ValueError

Both `init_checkpoint` and `model_config_file` should be spe

Error message

Both `init_checkpoint` and `model_config_file` should be specified if `config_file` is not specified.

What it means

Error "Both `init_checkpoint` and `model_config_file` should be specified if `config_file` is not specified." thrown in tensorflow/models.

Source

Thrown at official/nlp/finetuning/glue/flags.py:148

  if flags_obj.mode == 'predict':
    # model_dir is only needed strictly in 'predict' mode.
    _validate_path(flags_obj.model_dir, 'model_dir')

  if 'predict' in flags_obj.mode:
    _validate_path(flags_obj.test_input_path, 'test_input_path')

  if not flags_obj.config_file and flags_obj.mode != 'predict':
    if flags_obj.hub_module_url:
      if flags_obj.init_checkpoint or flags_obj.model_config_file:
        raise ValueError(
            'When `hub_module_url` is specified, `init_checkpoint` and '
            '`model_config_file` should be empty.')
      logging.info(
          'Using the pretrained tf.hub from %s', flags_obj.hub_module_url)
    else:
      if not (flags_obj.init_checkpoint and flags_obj.model_config_file):
        raise ValueError('Both `init_checkpoint` and `model_config_file` '
                         'should be specified if `config_file` is not '
                         'specified.')
      _validate_path(flags_obj.model_config_file, 'model_config_file')
      logging.info(
          'Using the pretrained checkpoint from %s and model_config_file from '
          '%s.', flags_obj.init_checkpoint, flags_obj.model_config_file)

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/nlp/finetuning/glue/flags.py:148 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of tensorflow/models@e006f5f0d5 (2026-08-24). Data as JSON: /api/errors/6b61b6baa21b55cf. Report an issue: GitHub.