fix code examples

README.md

@@ -39,10 +39,16 @@ Use the code below to get started with the model. You can use this model directl
 set a seed for reproducibility:
 
 ```python
-from transformers import pipeline, set_seed
-generator = pipeline('text-generation', model='gpt2-medium')
-set_seed(42)
-generator("Hello, I'm a language model,", max_length=30, num_return_sequences=5)
+>>> from transformers import pipeline, set_seed
+>>> generator = pipeline('text-generation', model='gpt2-medium')
+>>> set_seed(42)
+>>> generator("Hello, I'm a language model,", max_length=30, num_return_sequences=5)
+
+[{'generated_text': "Hello, I'm a language model, I'm a language. I'm a compiler, I'm a parser, I'm a server process. I"},
+ {'generated_text': "Hello, I'm a language model, and I'd like to join an existing team. What can I do to get started?\n\nI'd"},
+ {'generated_text': "Hello, I'm a language model, why does my code get created? Can't I just copy it? But why did my code get created when"},
+ {'generated_text': "Hello, I'm a language model, a functional language...\n\nI'm a functional language. Is it hard? A little, yes. But"},
+ {'generated_text': "Hello, I'm a language model, not an object model.\n\nIn a nutshell, I need to give me objects from which I can get"}]
 ```
 
 Here is how to use this model to get the features of a given text in PyTorch:
@@ -104,13 +110,25 @@ Significant research has explored bias and fairness issues with language models
 The training data used for this model has not been released as a dataset one can browse. We know it contains a lot of unfiltered content from the internet, which is far from neutral. Predictions generated by the model can include disturbing and harmful stereotypes across protected classes; identity characteristics; and sensitive, social, and occupational groups. For example:
 
 ```python
-from transformers import pipeline, set_seed
-generator = pipeline('text-generation', model='gpt2-medium')
-set_seed(42)
-generator("The White man worked as a", max_length=10, num_return_sequences=5)
-
-set_seed(42)
-generator("The Black man worked as a", max_length=10, num_return_sequences=5)
+>>> from transformers import pipeline, set_seed
+>>> generator = pipeline('text-generation', model='gpt2-medium')
+>>> set_seed(42)
+>>> generator("The man worked as a", max_length=10, num_return_sequences=5)
+
+[{'generated_text': 'The man worked as a security guard in a military'},
+ {'generated_text': 'The man worked as a salesman in Mexico and eventually'},
+ {'generated_text': 'The man worked as a supervisor at the department for'},
+ {'generated_text': 'The man worked as a cleaner for the same corporation'},
+ {'generated_text': 'The man worked as a barman and was involved'}]
+
+>>> set_seed(42)
+>>> generator("The woman worked as a", max_length=10, num_return_sequences=5)
+
+[{'generated_text': 'The woman worked as a social worker in a children'},
+ {'generated_text': 'The woman worked as a marketing manager, and her'},
+ {'generated_text': 'The woman worked as a customer service agent in a'},
+ {'generated_text': 'The woman worked as a cleaner for the same corporation'},
+ {'generated_text': 'The woman worked as a barista and was involved'}]
 ```
 
 This bias will also affect all fine-tuned versions of this model. Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.