bert-base for QA

Code: See Ainize Workspace

klue-bert-base-mrc DEMO: Ainize DEMO

klue-bert-base-mrc API: Ainize API


Language model: klue/bert-base
Language: Korean
Downstream-task: Extractive QA
Training data: KLUE-MRC
Eval data: KLUE-MRC


In Transformers

from transformers import AutoModelForQuestionAnswering, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("ainize/klue-bert-base-mrc")
model = AutoModelForQuestionAnswering.from_pretrained("ainize/klue-bert-base-mrc")

context = "your context"
question = "your question"

encodings = tokenizer(context, question, max_length=512, truncation=True,
                      padding="max_length", return_token_type_ids=False)
encodings = {key: torch.tensor([val]) for key, val in encodings.items()}             

input_ids = encodings["input_ids"]
attention_mask = encodings["attention_mask"]

pred = model(input_ids, attention_mask=attention_mask)

start_logits, end_logits = pred.start_logits, pred.end_logits

token_start_index, token_end_index = start_logits.argmax(dim=-1), end_logits.argmax(dim=-1)

pred_ids = input_ids[0][token_start_index: token_end_index + 1]

prediction = tokenizer.decode(pred_ids)

About us

Teachable NLP - Train NLP models with your own text without writing any code
Ainize - Deploy ML project using free gpu

Downloads last month
Hosted inference API
Question Answering
This model can be loaded on the Inference API on-demand.