dactyl-deberta-v3-large-pretrained

ShantanuT01/dactyl-deberta-v3-large-pretrained

dactyl-deberta-v3-large-pretrained is a text classification model from ShantanuT01. Use it when you need a label for a piece of text. It is set up for transformers.

Downloads · 30 days

6% of all-time downloads

All-time downloads

108

Public

Parameters

435M

1.7 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors1.7 GB · 100%

At a glance

Task: Text Classification
Library: transformers
Model type: deberta-v2
Access: Public
Created: May 28, 2025
Updated: Sep 24, 2025
SHA: c46aa3b9

Try a prompt

Model card

From the Hugging Face model README

DACTYL Text Detector

Training Configuration

{
        "training_split": "training",
        "evaluation_split": "testing",
        "results_path":"bce-debertav3-large.csv",
        "num_epochs": 1,
        "model_path": "best-debertav3-model",
        "tokenizer": "microsoft/deberta-v3-large",
        "optimizer": "AdamW",
        "optimizer_type": "torch",
        "optimizer_args": {
            "lr": 2e-5,
            "weight_decay":0.01
        },
        "loss_fn": "BCEWithLogitsLoss",
        "reset_classification_head": false,
        "loss_type": "torch",
        "loss_fn_args": {},
        "needs_loss_fn_as_parameter": false,
        "save_path": "ShantanuT01/dactyl-deberta-v3-large-pretrained",
        "training_args": {
            "batch_size": 16,
            "needs_sampler": false,
            "needs_index": false,
            "shuffle": true,
            "sampling_rate": null,
            "apply_sigmoid": false
        },
        "best_model_path": "best-debertav3-model",
        "eval_batch_size": 8
    
}

Results

model	AP Score	AUC Score	OPAUC Score	TPAUC Score
DeepSeek-V3	0.996018	0.999827	0.998268	0.98686
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-RedditWritingPrompts-testing	0.70532	0.959644	0.910133	0.160828
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-abstracts-testing	0.724459	0.980279	0.927757	0.411551
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-news-testing	0.42023	0.934212	0.840951	0
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-reviews-testing	0.320366	0.953743	0.834918	7.85083e-05
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-student_essays-testing	0.0152029	0.79227	0.510376	0
ShantanuT01/fine-tuned-Llama-3.2-1B-Instruct-apollo-mini-tweets-testing	0.468615	0.975752	0.856127	0.0610184
claude-3-5-haiku-20241022	0.972541	0.996902	0.982338	0.834824
claude-3-5-sonnet-20241022	0.992147	0.999509	0.995912	0.964305
gemini-1.5-flash	0.962676	0.995395	0.977143	0.784304
gemini-1.5-pro	0.919276	0.984604	0.949869	0.528626
gpt-4o-2024-11-20	0.973855	0.996913	0.983887	0.84817
gpt-4o-mini	0.994224	0.999781	0.998007	0.987244
llama-3.2-90b	0.938608	0.98916	0.960872	0.62814
llama-3.3-70b	0.977713	0.997508	0.986066	0.869837
mistral-large-latest	0.992386	0.999551	0.99628	0.968953
mistral-small-latest	0.993747	0.999556	0.997151	0.977661
overall	0.989294	0.991859	0.971624	0.730174

Task: Text Classification
Library: transformers
Type: deberta-v2
Created: May 28, 2025
Updated: Sep 24, 2025