Modelpublic

r2911/lm_p1

sourceHugging Facemitupdated 10mo agoView on Hugging Face

0likes9downloads

Model Card

🧠 AI Text Detector – DeBERTa v3 Large (Fine-tuned on Human vs AI Text)

This model is fine-tuned on a labeled dataset for AI-generated vs. Human-written text detection.

⚙️ Model Details

—Base Model: microsoft/deberta-v3-large
—Fine-tuned by: @abhinav
—Epochs: 4
—Learning Rate: 2e-05
—Batch Size: 8
—GPU: 80 GB A100
—Optimizer: AdamW (Fused)
—Scheduler: Cosine
—Mixed Precision: FP16
—Gradient Checkpointing: Enabled

📊 Evaluation Results (Test Set)

Metric	Score
Accuracy	0.993
Human (0) – Precision	1.000
Human (0) – Recall	0.986
AI (1) – Precision	0.988
AI (1) – Recall	1.000

🧮 Confusion Matrix

[image]

🚀 Example Inference

python

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tokenizer = AutoTokenizer.from_pretrained("abhi099k/ai-text-detector-deberta-v3-large-h1")
model = AutoModelForSequenceClassification.from_pretrained("abhi099k/ai-text-detector-deberta-v3-large-h1")

text = "This text was likely written by an AI model."
inputs = tokenizer(text, return_tensors="pt")
outputs = model(**inputs)
probs = torch.nn.functional.softmax(outputs.logits, dim=1)
print(probs)