Model Card for Hprophetnet-large
<!-- Provide a quick summary of what the model is/does. -->
This model is a fine-tuned version of all-mpnet-base-v2 on HP dataset to predict the popularity of the headlines. The input to the model is a headline and the ouput is a popularity score.
Intended uses & limitations
You can use this model to predict the popularity of the English-written headlines.
Usage
from transformers import AutoModelForSequenceClassification, AutoTokenizer
import torch
model = AutoModelForSequenceClassification.from_pretrained(
"omidvaramin/HeadlinePopularity", # We use the 12-layer BERT model, with an uncased vocab.
num_labels = 1, # The number of output labels--1 for regression.
output_attentions = False, # Whether the model returns attentions weights.
output_hidden_states = False, # Whether the model returns all hidden-states.
).cuda()
tokenizer = AutoTokenizer.from_pretrained("omidvaramin/HeadlinePopularity", do_lower_case=True)
HEADLINE = "Nigeria and Venezuela to Cut Oil Production"
encoded_dict = tokenizer.encode_plus(
HEADLINE, # Sentence to encode.
add_special_tokens = True, # Add '[CLS]' and '[SEP]'
max_length = 104, # Pad & truncate all sentences.
pad_to_max_length = True,
return_attention_mask = True, # Construct attn. masks.
return_tensors = 'pt', # Return pytorch tensors.
)
input_id = encoded_dict['input_ids']
input_id = input_id.cuda()
attention_mask = encoded_dict['attention_mask']
attention_mask = attention_mask.cuda()
with torch.no_grad():
result = model(input_id,
token_type_ids=None,
attention_mask=attention_mask,
return_dict=True)
popularity = result.logits.data[0].item()
print("Popularity: ", popularity)
>>> Popularity: 0.5718647837638855
BibTeX entry and citation info
@ARTICLE{10154027,
author={Omidvar, Amin and An, Aijun},
journal={IEEE Access},
title={Learning to Generate Popular Headlines},
year={2023},
volume={11},
number={},
pages={60904-60914},
doi={10.1109/ACCESS.2023.3286853}}