Merge pull request #866 from xanlsh/master

Rework how PreTrainedModel.from_pretrained handles its arguments

pytorch_transformers/modeling_utils.py

@@ -78,7 +78,7 @@ def save_pretrained(self, save_directory):
         self.to_json_file(output_config_file)
 
     @classmethod
-    def from_pretrained(cls, pretrained_model_name_or_path, *input, **kwargs):
+    def from_pretrained(cls, pretrained_model_name_or_path, **kwargs):
         r""" Instantiate a PretrainedConfig from a pre-trained model configuration.
 
         Params:
@@ -91,20 +91,33 @@ def from_pretrained(cls, pretrained_model_name_or_path, *input, **kwargs):
             **cache_dir**: (`optional`) string:
                 Path to a directory in which a downloaded pre-trained model
                 configuration should be cached if the standard cache should not be used.
+            **return_unused_kwargs**: (`optional`) bool:
+                - If False, then this function returns just the final configuration object.
+                - If True, then this functions returns a tuple `(config, unused_kwargs)` where `unused_kwargs`
+                is a dictionary consisting of the key/value pairs whose keys are not configuration attributes:
+                ie the part of kwargs which has not been used to update `config` and is otherwise ignored.
             **kwargs**: (`optional`) dict:
-                Dictionnary of key, values to update the configuration object after loading.
-                Can be used to override selected configuration parameters.
+                Dictionary of key/value pairs with which to update the configuration object after loading.
+                - The values in kwargs of any keys which are configuration attributes will be used
+                to override the loaded values.
+                - Behavior concerning key/value pairs whose keys are *not* configuration attributes is controlled
+                by the `return_unused_kwargs` keyword parameter.
 
         Examples::
 
             >>> config = BertConfig.from_pretrained('bert-base-uncased')    # Download configuration from S3 and cache.
             >>> config = BertConfig.from_pretrained('./test/saved_model/')  # E.g. config (or model) was saved using `save_pretrained('./test/saved_model/')`
             >>> config = BertConfig.from_pretrained('./test/saved_model/my_configuration.json')
-            >>> config = BertConfig.from_pretrained('bert-base-uncased', output_attention=True)
+            >>> config = BertConfig.from_pretrained('bert-base-uncased', output_attention=True, foo=False)
             >>> assert config.output_attention == True
+            >>> config, unused_kwargs = BertConfig.from_pretrained('bert-base-uncased', output_attention=True,
+            >>>                                                    foo=False, return_unused_kwargs=True)
+            >>> assert config.output_attention == True
+            >>> assert unused_kwargs == {'foo': False}
 
         """
         cache_dir = kwargs.pop('cache_dir', None)
+        return_unused_kwargs = kwargs.pop('return_unused_kwargs', False)
 
         if pretrained_model_name_or_path in cls.pretrained_config_archive_map:
             config_file = cls.pretrained_config_archive_map[pretrained_model_name_or_path]
@@ -148,7 +161,10 @@ def from_pretrained(cls, pretrained_model_name_or_path, *input, **kwargs):
             kwargs.pop(key, None)
 
         logger.info("Model config %s", config)
-        return config
+        if return_unused_kwargs:
+            return config, kwargs
+        else:
+            return config
 
     @classmethod
     def from_dict(cls, json_object):
@@ -305,7 +321,7 @@ def save_pretrained(self, save_directory):
         torch.save(model_to_save.state_dict(), output_model_file)
 
     @classmethod
-    def from_pretrained(cls, pretrained_model_name_or_path, *inputs, **kwargs):
+    def from_pretrained(cls, pretrained_model_name_or_path, *model_args, **kwargs):
         r"""Instantiate a pretrained pytorch model from a pre-trained model configuration.
 
             The model is set in evaluation mode by default using `model.eval()` (Dropout modules are desactivated)
@@ -322,6 +338,8 @@ def from_pretrained(cls, pretrained_model_name_or_path, *inputs, **kwargs):
                     provided as `config` argument. This loading option is slower than converting the TensorFlow
                     checkpoint in a PyTorch model using the provided conversion scripts and loading
                     the PyTorch model afterwards.
+            **model_args**: (`optional`) Sequence:
+                All remaning positional arguments will be passed to the underlying model's __init__ function
             **config**: an optional configuration for the model to use instead of an automatically loaded configuation.
                 Configuration can be automatically loaded when:
                 - the model is a model provided by the library (loaded with a `shortcut name` of a pre-trained model), or
@@ -337,8 +355,17 @@ def from_pretrained(cls, pretrained_model_name_or_path, *inputs, **kwargs):
             **output_loading_info**: (`optional`) boolean:
                 Set to ``True`` to also return a dictionnary containing missing keys, unexpected keys and error messages.
             **kwargs**: (`optional`) dict:
-                Dictionnary of key, values to update the configuration object after loading.
-                Can be used to override selected configuration parameters. E.g. ``output_attention=True``
+                Dictionary of key, values to update the configuration object after loading.
+                Can be used to override selected configuration parameters. E.g. ``output_attention=True``.
+
+               - If a configuration is provided with `config`, **kwargs will be directly passed
+                 to the underlying model's __init__ method.
+               - If a configuration is not provided, **kwargs will be first passed to the pretrained
+                 model configuration class loading function (`PretrainedConfig.from_pretrained`).
+                 Each key of **kwargs that corresponds to a configuration attribute
+                 will be used to override said attribute with the supplied **kwargs value.
+                 Remaining keys that do not correspond to any configuration attribute will
+                 be passed to the underlying model's __init__ function.
 
         Examples::
 
@@ -359,7 +386,13 @@ def from_pretrained(cls, pretrained_model_name_or_path, *inputs, **kwargs):
 
         # Load config
         if config is None:
-            config = cls.config_class.from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs)
+            config, model_kwargs = cls.config_class.from_pretrained(
+                pretrained_model_name_or_path, *model_args,
+                cache_dir=cache_dir, return_unused_kwargs=True,
+                **kwargs
+            )
+        else:
+            model_kwargs = kwargs
 
         # Load model
         if pretrained_model_name_or_path in cls.pretrained_model_archive_map:
@@ -400,7 +433,7 @@ def from_pretrained(cls, pretrained_model_name_or_path, *inputs, **kwargs):
                 archive_file, resolved_archive_file))
 
         # Instantiate model.
-        model = cls(config)
+        model = cls(config, *model_args, **model_kwargs)
 
         if state_dict is None and not from_tf:
             state_dict = torch.load(resolved_archive_file, map_location='cpu')
@@ -530,7 +563,7 @@ def forward(self, hidden_states, start_states=None, start_positions=None, p_mask
             **start_states**: ``torch.LongTensor`` of shape identical to hidden_states
                 hidden states of the first tokens for the labeled span.
             **start_positions**: ``torch.LongTensor`` of shape ``(batch_size,)``
-                position of the first token for the labeled span: 
+                position of the first token for the labeled span:
             **p_mask**: (`optional`) ``torch.FloatTensor`` of shape ``(batch_size, seq_len)``
                 Mask of invalid position such as query and special symbols (PAD, SEP, CLS)
                 1.0 means token should be masked.
@@ -717,7 +750,7 @@ class SequenceSummary(nn.Module):
                 - 'attn' => Not implemented now, use multi-head attention
             summary_use_proj: Add a projection after the vector extraction
             summary_proj_to_labels: If True, the projection outputs to config.num_labels classes (otherwise to hidden_size). Default: False.
-            summary_activation: 'tanh' => add a tanh activation to the output, Other => no activation. Default 
+            summary_activation: 'tanh' => add a tanh activation to the output, Other => no activation. Default
             summary_first_dropout: Add a dropout before the projection and activation
             summary_last_dropout: Add a dropout after the projection and activation
     """