Issue when running interact script

[harrystuart]

Hi there,

I am trying to run the interact script and am getting the following behaviour:

(logenv) C:\Users\Harry\source\repos\HuggingFace>python interact.py
2021-06-27 23:37:29.687306: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cudart64_101.dll
INFO:interact.py:Namespace(dataset_cache='./dataset_cache', dataset_path='', device='cpu', max_history=2, max_length=20, min_length=1, model='openai-gpt', model_checkpoint='', no_sample=False, seed=0, temperature=0.7, top_k=0, top_p=0.9)
INFO:C:\Users\Harry\source\repos\HuggingFace\utils.py:extracting archive file C:\Users\Harry/.cache\huggingface\transformers\2f5114b5eb72f9515802779c42c1b289bebdb1cfc8ce94c653237518eb530b34.75f2a4fe69178ff43138117a977e107a5fc7d402603a0825a296b531f246b3f2 to temp dir C:\Users\Harry\AppData\Local\Temp\tmpd181v50y
INFO:interact.py:Get pretrained model and tokenizer
ftfy or spacy is not installed using BERT BasicTokenizer instead of SpaCy & ftfy.
Some weights of the model checkpoint at C:\Users\Harry\AppData\Local\Temp\tmpd181v50y were not used when initializing OpenAIGPTLMHeadModel: ['multiple_choice_head.summary.bias', 'multiple_choice_head.summary.weight']
- This IS expected if you are initializing OpenAIGPTLMHeadModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).
- This IS NOT expected if you are initializing OpenAIGPTLMHeadModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).
INFO:interact.py:Sample a personality
INFO:C:\Users\Harry\source\repos\HuggingFace\utils.py:Load tokenized dataset from cache at ./dataset_cache_OpenAIGPTTokenizer
INFO:interact.py:Selected personality: i never have had alcohol in my life. i have a girlfriend of 7 years. i watch every football game at alabama. i currently suffer from social anxiety. i'm a geology major at alabama university.
>>> hey there
Traceback (most recent call last):
  File "interact.py", line 154, in <module>
    run()
  File "interact.py", line 146, in run
    out_ids = sample_sequence(personality, history, tokenizer, model, args)
  File "interact.py", line 72, in sample_sequence
    logits = logits[0, -1, :] / args.temperature
  File "C:\Users\Harry\Anaconda3\envs\logenv\lib\site-packages\transformers\file_utils.py", line 1808, in __getitem__
    return self.to_tuple()[k]
TypeError: tuple indices must be integers or slices, not tuple

(logenv) C:\Users\Harry\source\repos\HuggingFace>

Any ideas on a fix?
Thanks

[harrystuart]

Fix was to downgrade transformers version to that indicated in readme