Fix examples of loading pretrained models in docstring

pytorch_transformers/modeling_bert.py

@@ -643,12 +643,11 @@ class BertModel(BertPreTrainedModel):
 
     Examples::
 
-        config = BertConfig.from_pretrained('bert-base-uncased')
-        tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
-        model = BertModel(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
+        >>> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
+        >>> model = BertModel.from_pretrained('bert-base-uncased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
 
     """
     def __init__(self, config):
@@ -754,13 +753,11 @@ class BertForPreTraining(BertPreTrainedModel):
 
     Examples::
 
-        config = BertConfig.from_pretrained('bert-base-uncased')
-        tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
-        
-        model = BertForPreTraining(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        prediction_scores, seq_relationship_scores = outputs[:2]
+        >>> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
+        >>> model = BertForPreTraining.from_pretrained('bert-base-uncased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> prediction_scores, seq_relationship_scores = outputs[:2]
 
     """
     def __init__(self, config):
@@ -824,13 +821,11 @@ class BertForMaskedLM(BertPreTrainedModel):
 
     Examples::
 
-        config = BertConfig.from_pretrained('bert-base-uncased')
-        tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
-        
-        model = BertForMaskedLM(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, masked_lm_labels=input_ids)
-        loss, prediction_scores = outputs[:2]
+        >>> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
+        >>> model = BertForMaskedLM.from_pretrained('bert-base-uncased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, masked_lm_labels=input_ids)
+        >>> loss, prediction_scores = outputs[:2]
 
     """
     def __init__(self, config):
@@ -891,13 +886,11 @@ class BertForNextSentencePrediction(BertPreTrainedModel):
 
     Examples::
 
-        config = BertConfig.from_pretrained('bert-base-uncased')
-        tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
-        
-        model = BertForNextSentencePrediction(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        seq_relationship_scores = outputs[0]
+        >>> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
+        >>> model = BertForNextSentencePrediction.from_pretrained('bert-base-uncased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> seq_relationship_scores = outputs[0]
 
     """
     def __init__(self, config):
@@ -951,14 +944,12 @@ class BertForSequenceClassification(BertPreTrainedModel):
 
     Examples::
 
-        config = BertConfig.from_pretrained('bert-base-uncased')
-        tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
-        
-        model = BertForSequenceClassification(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, labels=labels)
-        loss, logits = outputs[:2]
+        >>> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
+        >>> model = BertForSequenceClassification.from_pretrained('bert-base-uncased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, labels=labels)
+        >>> loss, logits = outputs[:2]
 
     """
     def __init__(self, config):
@@ -1057,15 +1048,13 @@ class BertForMultipleChoice(BertPreTrainedModel):
 
     Examples::
 
-        config = BertConfig.from_pretrained('bert-base-uncased')
-        tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
-        
-        model = BertForMultipleChoice(config)
-        choices = ["Hello, my dog is cute", "Hello, my cat is amazing"]
-        input_ids = torch.tensor([tokenizer.encode(s) for s in choices]).unsqueeze(0)  # Batch size 1, 2 choices
-        labels = torch.tensor(1).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, labels=labels)
-        loss, classification_scores = outputs[:2]
+        >>> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
+        >>> model = BertForMultipleChoice.from_pretrained('bert-base-uncased')
+        >>> choices = ["Hello, my dog is cute", "Hello, my cat is amazing"]
+        >>> input_ids = torch.tensor([tokenizer.encode(s) for s in choices]).unsqueeze(0)  # Batch size 1, 2 choices
+        >>> labels = torch.tensor(1).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, labels=labels)
+        >>> loss, classification_scores = outputs[:2]
 
     """
     def __init__(self, config):
@@ -1127,14 +1116,12 @@ class BertForTokenClassification(BertPreTrainedModel):
 
     Examples::
 
-        config = BertConfig.from_pretrained('bert-base-uncased')
-        tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
-        
-        model = BertForTokenClassification(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        labels = torch.tensor([1] * input_ids.size(1)).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, labels=labels)
-        loss, scores = outputs[:2]
+        >>> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
+        >>> model = BertForTokenClassification.from_pretrained('bert-base-uncased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> labels = torch.tensor([1] * input_ids.size(1)).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, labels=labels)
+        >>> loss, scores = outputs[:2]
 
     """
     def __init__(self, config):
@@ -1203,15 +1190,13 @@ class BertForQuestionAnswering(BertPreTrainedModel):
 
     Examples::
 
-        config = BertConfig.from_pretrained('bert-base-uncased')
-        tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
-        
-        model = BertForQuestionAnswering(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        start_positions = torch.tensor([1])
-        end_positions = torch.tensor([3])
-        outputs = model(input_ids, start_positions=start_positions, end_positions=end_positions)
-        loss, start_scores, end_scores = outputs[:2]
+        >>> tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
+        >>> model = BertForQuestionAnswering.from_pretrained('bert-base-uncased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> start_positions = torch.tensor([1])
+        >>> end_positions = torch.tensor([3])
+        >>> outputs = model(input_ids, start_positions=start_positions, end_positions=end_positions)
+        >>> loss, start_scores, end_scores = outputs[:2]
 
     """
     def __init__(self, config):

pytorch_transformers/modeling_gpt2.py

@@ -433,12 +433,11 @@ class GPT2Model(GPT2PreTrainedModel):
 
     Examples::
 
-        config = GPT2Config.from_pretrained('gpt2')
-        tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
-        model = GPT2Model(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
+        >>> tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
+        >>> model = GPT2Model.from_pretrained('gpt2')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
 
     """
     def __init__(self, config):
@@ -567,12 +566,11 @@ class GPT2LMHeadModel(GPT2PreTrainedModel):
 
     Examples::
 
-        config = GPT2Config.from_pretrained('gpt2')
-        tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
-        model = GPT2LMHeadModel(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, labels=input_ids)
-        loss, logits = outputs[:2]
+        >>> tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
+        >>> model = GPT2LMHeadModel.from_pretrained('gpt2')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, labels=input_ids)
+        >>> loss, logits = outputs[:2]
 
     """
     def __init__(self, config):
@@ -683,14 +681,13 @@ class GPT2DoubleHeadsModel(GPT2PreTrainedModel):
 
     Examples::
 
-        config = GPT2Config.from_pretrained('gpt2')
-        tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
-        model = GPT2DoubleHeadsModel(config)
-        choices = ["Hello, my dog is cute [CLS]", "Hello, my cat is cute [CLS]"]  # Assume you've added [CLS] to the vocabulary
-        input_ids = torch.tensor([tokenizer.encode(s) for s in choices]).unsqueeze(0)  # Batch size 1, 2 choices
-        mc_token_ids = torch.tensor([-1, -1]).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, mc_token_ids)
-        lm_prediction_scores, mc_prediction_scores = outputs[:2]
+        >>> tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
+        >>> model = GPT2DoubleHeadsModel.from_pretrained('gpt2')
+        >>> choices = ["Hello, my dog is cute [CLS]", "Hello, my cat is cute [CLS]"]  # Assume you've added [CLS] to the vocabulary
+        >>> input_ids = torch.tensor([tokenizer.encode(s) for s in choices]).unsqueeze(0)  # Batch size 1, 2 choices
+        >>> mc_token_ids = torch.tensor([-1, -1]).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, mc_token_ids)
+        >>> lm_prediction_scores, mc_prediction_scores = outputs[:2]
 
     """
     def __init__(self, config):

pytorch_transformers/modeling_openai.py

@@ -439,12 +439,11 @@ class OpenAIGPTModel(OpenAIGPTPreTrainedModel):
 
     Examples::
 
-        config = OpenAIGPTConfig.from_pretrained('openai-gpt')
-        tokenizer = OpenAIGPTTokenizer.from_pretrained('openai-gpt')
-        model = OpenAIGPTModel(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
+        >>> tokenizer = OpenAIGPTTokenizer.from_pretrained('openai-gpt')
+        >>> model = OpenAIGPTModel.from_pretrained('openai-gpt')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
 
     """
     def __init__(self, config):
@@ -558,12 +557,11 @@ class OpenAIGPTLMHeadModel(OpenAIGPTPreTrainedModel):
 
     Examples::
 
-        config = OpenAIGPTConfig.from_pretrained('openai-gpt')
-        tokenizer = OpenAIGPTTokenizer.from_pretrained('openai-gpt')
-        model = OpenAIGPTLMHeadModel(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, labels=input_ids)
-        loss, logits = outputs[:2]
+        >>> tokenizer = OpenAIGPTTokenizer.from_pretrained('openai-gpt')
+        >>> model = OpenAIGPTLMHeadModel.from_pretrained('openai-gpt')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, labels=input_ids)
+        >>> loss, logits = outputs[:2]
 
     """
     def __init__(self, config):
@@ -665,14 +663,13 @@ class OpenAIGPTDoubleHeadsModel(OpenAIGPTPreTrainedModel):
 
     Examples::
 
-        config = OpenAIGPTConfig.from_pretrained('openai-gpt')
-        tokenizer = OpenAIGPTTokenizer.from_pretrained('openai-gpt')
-        model = OpenAIGPTDoubleHeadsModel(config)
-        choices = ["Hello, my dog is cute [CLS]", "Hello, my cat is cute [CLS]"]  # Assume you've added [CLS] to the vocabulary
-        input_ids = torch.tensor([tokenizer.encode(s) for s in choices]).unsqueeze(0)  # Batch size 1, 2 choices
-        mc_token_ids = torch.tensor([-1, -1]).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, mc_token_ids)
-        lm_prediction_scores, mc_prediction_scores = outputs[:2]
+        >>> tokenizer = OpenAIGPTTokenizer.from_pretrained('openai-gpt')
+        >>> model = OpenAIGPTDoubleHeadsModel.from_pretrained('openai-gpt')
+        >>> choices = ["Hello, my dog is cute [CLS]", "Hello, my cat is cute [CLS]"]  # Assume you've added [CLS] to the vocabulary
+        >>> input_ids = torch.tensor([tokenizer.encode(s) for s in choices]).unsqueeze(0)  # Batch size 1, 2 choices
+        >>> mc_token_ids = torch.tensor([-1, -1]).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, mc_token_ids)
+        >>> lm_prediction_scores, mc_prediction_scores = outputs[:2]
 
     """
     def __init__(self, config):

pytorch_transformers/modeling_transfo_xl.py

@@ -968,12 +968,11 @@ class TransfoXLModel(TransfoXLPreTrainedModel):
 
     Examples::
 
-        config = TransfoXLConfig.from_pretrained('transfo-xl-wt103')
-        tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
-        model = TransfoXLModel(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        last_hidden_states, mems = outputs[:2]
+        >>> tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
+        >>> model = TransfoXLModel.from_pretrained('transfo-xl-wt103')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> last_hidden_states, mems = outputs[:2]
 
     """
     def __init__(self, config):
@@ -1284,12 +1283,11 @@ class TransfoXLLMHeadModel(TransfoXLPreTrainedModel):
 
     Examples::
 
-        config = TransfoXLConfig.from_pretrained('transfo-xl-wt103')
-        tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
-        model = TransfoXLLMHeadModel(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        prediction_scores, mems = outputs[:2]
+        >>> tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
+        >>> model = TransfoXLLMHeadModel.from_pretrained('transfo-xl-wt103')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> prediction_scores, mems = outputs[:2]
 
     """
     def __init__(self, config):

pytorch_transformers/modeling_xlm.py

@@ -472,12 +472,11 @@ class XLMModel(XLMPreTrainedModel):
 
     Examples::
 
-        config = XLMConfig.from_pretrained('xlm-mlm-en-2048')
-        tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
-        model = XLMModel(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
+        >>> tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
+        >>> model = XLMModel.from_pretrained('xlm-mlm-en-2048')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
 
     """
     ATTRIBUTES = ['encoder', 'eos_index', 'pad_index',  # 'with_output', 
@@ -745,12 +744,11 @@ class XLMWithLMHeadModel(XLMPreTrainedModel):
 
     Examples::
 
-        config = XLMConfig.from_pretrained('xlm-mlm-en-2048')
-        tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
-        model = XLMWithLMHeadModel(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
+        >>> tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
+        >>> model = XLMWithLMHeadModel.from_pretrained('xlm-mlm-en-2048')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
 
     """
     def __init__(self, config):
@@ -805,14 +803,12 @@ class XLMForSequenceClassification(XLMPreTrainedModel):
 
     Examples::
 
-        config = XLMConfig.from_pretrained('xlm-mlm-en-2048')
-        tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
-        
-        model = XLMForSequenceClassification(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, labels=labels)
-        loss, logits = outputs[:2]
+        >>> tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
+        >>> model = XLMForSequenceClassification.from_pretrained('xlm-mlm-en-2048')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, labels=labels)
+        >>> loss, logits = outputs[:2]
 
     """
     def __init__(self, config):
@@ -885,15 +881,13 @@ class XLMForQuestionAnswering(XLMPreTrainedModel):
 
     Examples::
 
-        config = XLMConfig.from_pretrained('xlm-mlm-en-2048')
-        tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
-        
-        model = XLMForQuestionAnswering(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        start_positions = torch.tensor([1])
-        end_positions = torch.tensor([3])
-        outputs = model(input_ids, start_positions=start_positions, end_positions=end_positions)
-        loss, start_scores, end_scores = outputs[:2]
+        >>> tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
+        >>> model = XLMForQuestionAnswering.from_pretrained('xlm-mlm-en-2048')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> start_positions = torch.tensor([1])
+        >>> end_positions = torch.tensor([3])
+        >>> outputs = model(input_ids, start_positions=start_positions, end_positions=end_positions)
+        >>> loss, start_scores, end_scores = outputs[:2]
 
     """
     def __init__(self, config):

pytorch_transformers/modeling_xlnet.py

@@ -712,12 +712,11 @@ class XLNetModel(XLNetPreTrainedModel):
 
     Examples::
 
-        config = XLNetConfig.from_pretrained('xlnet-large-cased')
-        tokenizer = XLNetTokenizer.from_pretrained('xlnet-large-cased')
-        model = XLNetModel(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids)
-        last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
+        >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-large-cased')
+        >>> model = XLNetModel.from_pretrained('xlnet-large-cased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids)
+        >>> last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
 
     """
     def __init__(self, config):
@@ -1019,17 +1018,16 @@ class XLNetLMHeadModel(XLNetPreTrainedModel):
 
     Examples::
 
-        config = XLNetConfig.from_pretrained('xlnet-large-cased')
-        tokenizer = XLNetTokenizer.from_pretrained('xlnet-large-cased')
-        model = XLNetLMHeadModel(config)
-        # We show how to setup inputs to predict a next token using a bi-directional context.
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is very <mask>")).unsqueeze(0)  # We will predict the masked token
-        perm_mask = torch.zeros((1, input_ids.shape[1], input_ids.shape[1]), dtype=torch.float)
-        perm_mask[:, :, -1] = 1.0  # Previous tokens don't see last token
-        target_mapping = torch.zeros((1, 1, input_ids.shape[1]), dtype=torch.float)  # Shape [1, 1, seq_length] => let's predict one token
-        target_mapping[0, 0, -1] = 1.0  # Our first (and only) prediction will be the last token of the sequence (the masked token)
-        outputs = model(input_ids, perm_mask=perm_mask, target_mapping=target_mapping)
-        next_token_logits = outputs[0]  # Output has shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
+        >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-large-cased')
+        >>> model = XLNetLMHeadModel.from_pretrained('xlnet-large-cased')
+        >>> # We show how to setup inputs to predict a next token using a bi-directional context.
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is very <mask>")).unsqueeze(0)  # We will predict the masked token
+        >>> perm_mask = torch.zeros((1, input_ids.shape[1], input_ids.shape[1]), dtype=torch.float)
+        >>> perm_mask[:, :, -1] = 1.0  # Previous tokens don't see last token
+        >>> target_mapping = torch.zeros((1, 1, input_ids.shape[1]), dtype=torch.float)  # Shape [1, 1, seq_length] => let's predict one token
+        >>> target_mapping[0, 0, -1] = 1.0  # Our first (and only) prediction will be the last token of the sequence (the masked token)
+        >>> outputs = model(input_ids, perm_mask=perm_mask, target_mapping=target_mapping)
+        >>> next_token_logits = outputs[0]  # Output has shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
 
     """
     def __init__(self, config):
@@ -1100,14 +1098,12 @@ class XLNetForSequenceClassification(XLNetPreTrainedModel):
 
     Examples::
 
-        config = XLNetConfig.from_pretrained('xlnet-large-cased')
-        tokenizer = XLNetTokenizer.from_pretrained('xlnet-large-cased')
-        
-        model = XLNetForSequenceClassification(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
-        outputs = model(input_ids, labels=labels)
-        loss, logits = outputs[:2]
+        >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-large-cased')
+        >>> model = XLNetForSequenceClassification.from_pretrained('xlnet-large-cased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
+        >>> outputs = model(input_ids, labels=labels)
+        >>> loss, logits = outputs[:2]
 
     """
     def __init__(self, config):
@@ -1200,15 +1196,13 @@ class XLNetForQuestionAnswering(XLNetPreTrainedModel):
 
     Examples::
 
-        config = XLMConfig.from_pretrained('xlm-mlm-en-2048')
-        tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
-        
-        model = XLMForQuestionAnswering(config)
-        input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
-        start_positions = torch.tensor([1])
-        end_positions = torch.tensor([3])
-        outputs = model(input_ids, start_positions=start_positions, end_positions=end_positions)
-        loss, start_scores, end_scores = outputs[:2]
+        >>> tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
+        >>> model = XLMForQuestionAnswering.from_pretrained('xlnet-large-cased')
+        >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute")).unsqueeze(0)  # Batch size 1
+        >>> start_positions = torch.tensor([1])
+        >>> end_positions = torch.tensor([3])
+        >>> outputs = model(input_ids, start_positions=start_positions, end_positions=end_positions)
+        >>> loss, start_scores, end_scores = outputs[:2]
 
     """
     def __init__(self, config):