ai-text-detector-deberta-v3-large-80-gb

This model is a fine-tuned version of microsoft/deberta-v3-large on an unknown dataset.

Model description

MODEL_REPO = "abhi099k/ai-text-detector-deberta-v3-large-80-gb" MODEL_NAME = "AI Text Detector - DeBERTa v3 Large (80GB GPU)" BASE_MODEL = "microsoft/deberta-v3-large"

🧠 ai-text-detector-deberta-v3-large-80-gb

This model is fine-tuned on a Human vs AI Text Detection dataset using the base model:
microsoft/deberta-v3-large

---

📊 Training Configuration

Parameter	Value
Epochs	4
Batch Size	8
Learning Rate	2e-5
Weight Decay	0.01
GPU Used	80GB A100
Framework	🤗 Transformers + PyTorch

---

📈 Evaluation Results (Test Set)

Metric	Score
Accuracy	0.993
F1 Score	0.994

🔹 Class-wise Performance

Class	Precision	Recall	F1-score
Human (0)	1.000	0.986	0.993
AI (1)	0.988	1.000	0.994

{conf_matrix_section}

---

🚀 Usage Example

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tokenizer = AutoTokenizer.from_pretrained("{MODEL_REPO}")
model = AutoModelForSequenceClassification.from_pretrained("{MODEL_REPO}")

text = "This text might have been written by an AI."
inputs = tokenizer(text, return_tensors="pt")
outputs = model(**inputs)

probs = torch.nn.functional.softmax(outputs.logits, dim=1)
labels = ["Human", "AI"]
predicted_label = labels[torch.argmax(probs)]
print(predicted_label, probs)