Model Description

This model detects if you are writing in a format that is more similar to Simple English Wikipedia or English Wikipedia. This can be extended to applications that aren't Wikipedia as well and to some extent, it can be used for other languages.

Please also note there is a major bias to special characters (Mainly the hyphen mark, but it also applies to others) so I would recommend removing them from your input text.

Model Trained Using AutoTrain

• Problem type: Multi-class Classification
• Model ID: 837726721
• CO2 Emissions (in grams): 0.21691606119445225

Validation Metrics

• Loss: 0.010096958838403225
• Accuracy: 0.996223414828066
• Macro F1: 0.996179398826373
• Micro F1: 0.996223414828066
• Weighted F1: 0.996223414828066
• Macro Precision: 0.996179398826373
• Micro Precision: 0.996223414828066
• Weighted Precision: 0.996223414828066
• Macro Recall: 0.996179398826373
• Micro Recall: 0.996223414828066
• Weighted Recall: 0.996223414828066

Usage

You can use cURL to access this model:

$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I quite enjoy using AutoTrain due to its simplicity."}' https://api-inference.huggingface.co/models/hidude562/Wiki-Complexity

Or Python API:

from transformers import AutoModelForSequenceClassification, AutoTokenizer

model = AutoModelForSequenceClassification.from_pretrained("hidude562/Wiki-Complexity", use_auth_token=True)

tokenizer = AutoTokenizer.from_pretrained("hidude562/Wiki-Complexity", use_auth_token=True)

inputs = tokenizer("I quite enjoy using AutoTrain due to its simplicity.", return_tensors="pt")

outputs = model(**inputs)