melanchthon19/mews-aes-regressor-base

DistilBERT Regressor (Test Model)

This repository contains a DistilBERT-based regression model with a lightweight feed-forward head trained for essay scoring.

The model combines a frozen distilbert-base-uncased encoder with a two-layer MLP regression head.
It was trained for 1 epoch purely for testing and demonstration purposes — not intended for production or research use.

Usage

from transformers import AutoTokenizer, AutoModel

model_name = "melanchthon19/mews-aes-regressor-base"
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModel.from_pretrained(model_name, trust_remote_code=True)

text = "Artificial intelligence models are transforming how we write and analyze essays."
inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
pred = model(**inputs)
print(pred.item())