Spaces:
Running
Running
from transformers import pipeline | |
# import transformers | |
# import torch | |
# import torch.nn as nn | |
# import tensorflow as tf | |
# from transformers import TFGPT2LMHeadModel ,GPT2Tokenizer, BitsAndBytesConfig | |
# tokenizer = GPT2Tokenizer.from_pretrained('gpt2') | |
# model = TFGPT2LMHeadModel.from_pretrained('gpt2',pad_token_id = tokenizer.eos_token_id) | |
# def generate_text(inp): | |
# input_ids = tokenizer.encode(inp,return_tensors = 'tf') | |
# beam_output = model.generate(input_ids, max_length = 100,num_beams = 5, no_repeat_ngram_size = 2, early_stopping = True) | |
# output = tokenizer.decode(beam_output[0],skip_special_tokens = True, clean_up_tokenization_spaces = True) | |
# return ".".join(output.split(".")[:-1]) + "." | |
qa_pipeline = pipeline("question-answering", model='deepset/roberta-base-squad2') |