BART-LARGE finetuned on SQuADv1
This is bart-large model finetuned on SQuADv1 dataset for question answering task
Model details
BART was propsed in the paper BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension. BART is a seq2seq model intended for both NLG and NLU tasks.
To use BART for question answering tasks, we feed the complete document into the encoder and decoder, and use the top hidden state of the decoder as a representation for each word. This representation is used to classify the token. As given in the paper bart-large achives comparable to ROBERTa on SQuAD. Another notable thing about BART is that it can handle sequences with upto 1024 tokens.
Param | #Value |
---|---|
encoder layers | 12 |
decoder layers | 12 |
hidden size | 4096 |
num attetion heads | 16 |
on disk size | 1.63GB |
Model training
This model was trained on google colab v100 GPU. You can find the fine-tuning colab here .
Results
The results are actually slightly worse than given in the paper. In the paper the authors mentioned that bart-large achieves 88.8 EM and 94.6 F1
Metric | #Value |
---|---|
EM | 86.8022 |
F1 | 92.7342 |
Model in Action 🚀
from transformers import BartTokenizer, BartForQuestionAnswering
import torch
tokenizer = BartTokenizer.from_pretrained('valhalla/bart-large-finetuned-squadv1')
model = BartForQuestionAnswering.from_pretrained('valhalla/bart-large-finetuned-squadv1')
question, text = "Who was Jim Henson?", "Jim Henson was a nice puppet"
encoding = tokenizer(question, text, return_tensors='pt')
input_ids = encoding['input_ids']
attention_mask = encoding['attention_mask']
start_scores, end_scores = model(input_ids, attention_mask=attention_mask, output_attentions=False)[:2]
all_tokens = tokenizer.convert_ids_to_tokens(input_ids[0])
answer = ' '.join(all_tokens[torch.argmax(start_scores) : torch.argmax(end_scores)+1])
answer = tokenizer.convert_tokens_to_ids(answer.split())
answer = tokenizer.decode(answer)
#answer => 'a nice puppet'
- Downloads last month
- 1,079