Model Trained Using AutoNLP

Problem type: Extractive Question Answering
Model ID: 26265903
CO2 Emissions (in grams): 108.13983395548236

Validation Metrics

Loss: 0.6330059170722961

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_24465520-26265903

Or Python API:

import torch

from transformers import AutoModelForQuestionAnswering, AutoTokenizer

model = AutoModelForQuestionAnswering.from_pretrained("teacookies/autonlp-more_fine_tune_24465520-26265903", use_auth_token=True)

tokenizer = AutoTokenizer.from_pretrained("teacookies/autonlp-more_fine_tune_24465520-26265903", 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