CTRL model can not work in huggingface transformers

[yanan1116]

torch ver: 1.8.1+cu102 transformers ver: 4.4.2

I adopt the example codes from https://github.com/huggingface/transformers/blob/master/examples/text-generation/run_generation.py to generate text by using ctrl. here is the head part of my codes:

import torch
from transformers import CTRLTokenizer, CTRLLMHeadModel
tokenizer = CTRLTokenizer.from_pretrained('ctrl')
model = CTRLLMHeadModel.from_pretrained('ctrl')

encoded_prompt = tokenizer.encode("Links Hello, my dog is cute", add_special_tokens=False)

ERROR:

ValueError Traceback (most recent call last) in ----> 1 encoded_prompt = tokenizer.encode("Links Hello, my dog is cute", add_special_tokens=False)

~/yanan/env/lib/python3.6/site-packages/transformers/tokenization_utils_base.py in encode(self, text, text_pair, add_special_tokens, padding, truncation, max_length, stride, return_tensors, **kwargs) 2030 stride=stride, 2031 return_tensors=return_tensors, -> 2032 **kwargs, 2033 ) 2034

~/yanan/env/lib/python3.6/site-packages/transformers/tokenization_utils_base.py in encode_plus(self, text, text_pair, add_special_tokens, padding, truncation, max_length, stride, is_split_into_words, pad_to_multiple_of, return_tensors, return_token_type_ids, return_attention_mask, return_overflowing_tokens, return_special_tokens_mask, return_offsets_mapping, return_length, verbose, **kwargs) 2355 return_length=return_length, 2356 verbose=verbose, -> 2357 **kwargs, 2358 ) 2359

~/yanan/env/lib/python3.6/site-packages/transformers/tokenization_utils.py in _encode_plus(self, text, text_pair, add_special_tokens, padding_strategy, truncation_strategy, max_length, stride, is_split_into_words, pad_to_multiple_of, return_tensors, return_token_type_ids, return_attention_mask, return_overflowing_tokens, return_special_tokens_mask, return_offsets_mapping, return_length, verbose, **kwargs) 458 return_special_tokens_mask=return_special_tokens_mask, 459 return_length=return_length, --> 460 verbose=verbose, 461 ) 462

~/yanan/env/lib/python3.6/site-packages/transformers/tokenization_utils_base.py in prepare_for_model(self, ids, pair_ids, add_special_tokens, padding, truncation, max_length, stride, pad_to_multiple_of, return_tensors, return_token_type_ids, return_attention_mask, return_overflowing_tokens, return_special_tokens_mask, return_offsets_mapping, return_length, verbose, prepend_batch_axis, **kwargs) 2792 padding=padding_strategy.value, 2793 pad_to_multiple_of=pad_to_multiple_of, -> 2794 return_attention_mask=return_attention_mask, 2795 ) 2796

~/yanan/env/lib/python3.6/site-packages/transformers/tokenization_utils_base.py in pad(self, encoded_inputs, padding, max_length, pad_to_multiple_of, return_attention_mask, return_tensors, verbose) 2591 else: 2592 raise ValueError( -> 2593 f"type of {first_element} unknown: {type(first_element)}. " 2594 f"Should be one of a python, numpy, pytorch or tensorflow object." 2595 )

ValueError: type of None unknown: <class 'NoneType'>. Should be one of a python, numpy, pytorch or tensorflow object.

[yanan1116]

python run_generation.py --model_type ctrl --model_name ctrl --temperature 0 --repetition 1.2

04/07/2021 10:22:08 - WARNING - main - device: cpu, n_gpu: 0, 16-bits training: False 04/07/2021 10:23:25 - INFO - main - Namespace(device=device(type='cpu'), fp16=False, k=0, length=20, model_name_or_path='ctrl', model_type='ctrl', n_gpu=0, no_cuda=False, num_return_sequences=1, p=0.9, padding_text='', prefix='', prompt='', repetition_penalty=1.2, seed=42, stop_token=None, temperature=0.0, xlm_language='') Model prompt >>> Links Hello, my dog is cute 2021-04-07 10:24:00.280690: W tensorflow/stream_executor/platform/default/dso_loader.cc:60] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory 2021-04-07 10:24:00.280732: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine. Traceback (most recent call last): File "run_generation.py", line 290, in main() File "run_generation.py", line 231, in main preprocessed_prompt_text = prepare_input(args, model, tokenizer, prompt_text) File "run_generation.py", line 92, in prepare_ctrl_input encoded_prompt = tokenizer.encode(prompt_text, add_special_tokens=False) File "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py", line 2032, in encode **kwargs, File "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py", line 2357, in encode_plus **kwargs, File "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils.py", line 460, in _encode_plus verbose=verbose, File "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py", line 2794, in prepare_for_model return_attention_mask=return_attention_mask, File "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py", line 2593, in pad f"type of {first_element} unknown: {type(first_element)}. " ValueError: type of None unknown: <class 'NoneType'>. Should be one of a python, numpy, pytorch or tensorflow object.

[sandcobainer]

Any updates on this? I haven't found any documentation about how to run the CTRL model from huggingface transformers either.