bharathja/tfplan-risk-classifier-deberta-v3-small

Hugging Face bundle: Terraform Plan Risk Labels + Baseline Classifier

This bundle contains three Hugging Face‑ready repos:

1. Dataset: bharathja/tfplan-risk-labels-multicloud
2. Model: bharathja/tfplan-risk-classifier-deberta-v3-small
3. (Optional) CLI helper: bharathja/tfplan-risk-cli

---

Publish order (recommended)

1) Create the dataset repo on the Hub

Create a Dataset repo named: tfplan-risk-labels-multicloud under your account, then push the contents of tfplan-risk-labels/.

cd tfplan-risk-labels
git init
git lfs install
git remote add origin https://huggingface.co/datasets/bharathja/tfplan-risk-labels-multicloud
git add .
git commit -m "init dataset"
git push -u origin main

2) Train the baseline classifier locally

cd ../tfplan-risk-classifier-deberta-v3-small
python -m pip install -r requirements.txt
python train.py --dataset_path ../tfplan-risk-labels --output_dir ./out --epochs 3

3) Create the model repo and push the trained artifacts

Create a Model repo named: tfplan-risk-classifier-deberta-v3-small, then push out/.

cd out
git init
git lfs install
git remote add origin https://huggingface.co/bharathja/tfplan-risk-classifier-deberta-v3-small
cp ../README.md ./README.md
git add .
git commit -m "add model"
git push -u origin main

---

Quick sanity check (after publish)

Load dataset

from datasets import load_dataset
ds = load_dataset("bharathja/tfplan-risk-labels-multicloud", data_files={"train":"train.jsonl","validation":"validation.jsonl","test":"test.jsonl"})
print(ds["train"][0]["labels"])

Load model

from transformers import AutoTokenizer, AutoModelForSequenceClassification
tok = AutoTokenizer.from_pretrained("bharathja/tfplan-risk-classifier-deberta-v3-small")
model = AutoModelForSequenceClassification.from_pretrained("bharathja/tfplan-risk-classifier-deberta-v3-small")