def create_and_check_bert_for_question_answering(
self,
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
):
model = BertForQuestionAnswering(config=config)
model.eval()
loss, start_logits, end_logits = model(input_ids, token_type_ids,
input_mask, sequence_labels,
sequence_labels)
result = {
"loss": loss,
"start_logits": start_logits,
"end_logits": end_logits
}
self.parent.assertListEqual(list(result["start_logits"].size()),
[self.batch_size, self.seq_length])
self.parent.assertListEqual(list(result["end_logits"].size()),
[self.batch_size, self.seq_length])
self.check_loss_output(result)
def __init__(self, model_state_dict) -> None:
no_cuda = True
self.device = torch.device(
"cuda" if torch.cuda.is_available() and not no_cuda else "cpu")
self.tokenizer = BertTokenizer.from_pretrained('bert-base-chinese',
do_lower_case=False)
config = BertConfig.from_pretrained('bert-base-chinese')
self.model = BertForQuestionAnswering(config)
self.model.load_state_dict(
torch.load(model_state_dict, map_location='cpu'))
self.model.to(self.device)
self.model.eval() # TODO
type=str,
required=True,
help="model para after pretrained")
args = parser.parse_args()
args.n_gpu = torch.cuda.device_count()
args.eval_batch_size = args.per_gpu_eval_batch_size * max(1, args.n_gpu)
device = torch.device(
"cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu")
args.device = device
tokenizer = BertTokenizer.from_pretrained(
args.tokenizer_name if args.tokenizer_name else args.model_name_or_path,
do_lower_case=False)
config = BertConfig.from_pretrained(
args.tokenizer_name if args.tokenizer_name else args.model_name_or_path)
model = BertForQuestionAnswering(config)
model_state_dict = args.state_dict
model.load_state_dict(torch.load(model_state_dict))
model.to(args.device)
model.eval()
input_file = args.predict_file
def handle_file(input_file, context, question):
with open(input_file, "r") as reader:
orig_data = json.load(reader)
orig_data["data"][0]['paragraphs'][0]['context'] = context
for i in range(len(question)):
orig_data["data"][0]['paragraphs'][0]['qas'][i][
'question'] = question[i]
with open(input_file, "w") as writer: