abhishek's picture
abhishek HF staff
Commit From AutoNLP
bb23248
metadata
tags:
  - autonlp
  - question-answering
language: unk
widget:
  - text: Who loves AutoNLP?
    context: Everyone loves AutoNLP
datasets:
  - teacookies/autonlp-data-more_fine_tune_24465520
co2_eq_emissions: 83.00580438705762

Model Trained Using AutoNLP

  • Problem type: Extractive Question Answering
  • Model ID: 26265906
  • CO2 Emissions (in grams): 83.00580438705762

Validation Metrics

  • Loss: 0.5259918570518494

Usage

You can use cURL to access this model:

$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"question": "Who loves AutoNLP?", "context": "Everyone loves AutoNLP"}' https://api-inference.huggingface.co/models/teacookies/autonlp-more_fine_tune_24465520-26265906

Or Python API:

import torch

from transformers import AutoModelForQuestionAnswering, AutoTokenizer

model = AutoModelForQuestionAnswering.from_pretrained("teacookies/autonlp-more_fine_tune_24465520-26265906", use_auth_token=True)

tokenizer = AutoTokenizer.from_pretrained("teacookies/autonlp-more_fine_tune_24465520-26265906", use_auth_token=True)

from transformers import BertTokenizer, BertForQuestionAnswering

question, text = "Who loves AutoNLP?", "Everyone loves AutoNLP"

inputs = tokenizer(question, text, return_tensors='pt')

start_positions = torch.tensor([1])

end_positions = torch.tensor([3])

outputs = model(**inputs, start_positions=start_positions, end_positions=end_positions)

loss = outputs.loss

start_scores = outputs.start_logits

end_scores = outputs.end_logits