Pytorch example encoded_input fix

There is a problem like in t5-large from hugging face" with 'size'. Link to solved issue https://github.com/huggingface/transformers/issues/5480.
The returned object is a dict, not a tensor for the model. ['input_ids'] fixes PyTorch example.

README.md

@@ -49,7 +49,7 @@ Here is how to use this model to get the features of a given text in PyTorch:
 tokenizer = BertTokenizer.from_pretrained('tbs17/MathBERT-custom')
 model = BertModel.from_pretrained("tbs17/MathBERT-custom")
 text = "Replace me by any text you'd like."
-encoded_input = tokenizer(text, return_tensors='pt')
+encoded_input = tokenizer(text, return_tensors='pt')["input_ids"]
 output = model(encoded_input)
 ```
 and in TensorFlow: