Supply Chain DistilBERT Risk Classifier

Fine-tuned distilbert-base-uncased for 4-class electronics supply chain risk: LOW | MEDIUM | HIGH | CRITICAL

Trained on lite_master rows from the supply-chain-disrupter project.

Usage

from transformers import AutoModelForSequenceClassification, AutoTokenizer
import torch

model_id = "mathurvarun84/supply-chain-distilbert-risk"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForSequenceClassification.from_pretrained(model_id)

text = "Region: Eastern Asia. Product: DRAM. Delivery: Late delivery. ..."
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=256)
with torch.no_grad():
    pred = model(**inputs).logits.argmax(-1).item()
labels = ["LOW", "MEDIUM", "HIGH", "CRITICAL"]
print(labels[pred])

Local project

Set in .env:

DISTILBERT_MODEL_ID=mathurvarun84/supply-chain-distilbert-risk