Encoder decoder config docs (#6195)

* Adding docs for how to load encoder_decoder pretrained model with individual config objects

* Adding docs for loading encoder_decoder config from pretrained folder

* Fixing  W293 blank line contains whitespace

* Update src/transformers/modeling_encoder_decoder.py

* Update src/transformers/modeling_encoder_decoder.py

* Update src/transformers/modeling_encoder_decoder.py

* Apply suggestions from code review

model file should only show examples for how to load save model

* Update src/transformers/configuration_encoder_decoder.py

* Update src/transformers/configuration_encoder_decoder.py

* fix space

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

src/transformers/configuration_encoder_decoder.py

@@ -56,6 +56,15 @@ class EncoderDecoderConfig(PretrainedConfig):
             >>> # Accessing the model configuration
             >>> config_encoder = model.config.encoder
             >>> config_decoder  = model.config.decoder
+            >>> # set decoder config to causal lm
+            >>> config_decoder.is_decoder = True
+
+            >>> # Saving the model, including its configuration
+            >>> model.save_pretrained('my-model')
+
+            >>> # loading model and config from pretrained folder
+            >>> encoder_decoder_config = EncoderDecoderConfig.from_pretrained('my-model')
+            >>> model = EncoderDecoderModel.from_pretrained('my-model', config=encoder_decoder_config)
     """
     model_type = "encoder_decoder"
 

src/transformers/modeling_encoder_decoder.py

@@ -127,7 +127,13 @@ def from_encoder_decoder_pretrained(
         Examples::
 
             >>> from transformers import EncoderDecoderModel
-            >>> model = EncoderDecoderModel.from_encoder_decoder_pretrained('bert-base-uncased', 'bert-base-uncased') # initialize Bert2Bert
+            >>> # initialize a bert2bert from two pretrained BERT models. Note that the cross-attention layers will be randomly initialized
+            >>> model = EncoderDecoderModel.from_encoder_decoder_pretrained('bert-base-uncased', 'bert-base-uncased')
+            >>> # saving model after fine-tuning
+            >>> model.save_pretrained("./bert2bert")
+            >>> # load fine-tuned model
+            >>> model = EncoderDecoderModel.from_pretrained("./bert2bert")
+
         """
 
         kwargs_encoder = {