Skip to content
Other

AI image detector v5

AI image detector

2 min readupdated

This model is a research preview image classification kit designed to locally detect AI-generated content directly in the browser. It uses a custom CLIP feature extractor (a powerful model for matching text and images) and a calibrated linear ensemble. The tool outputs score signals for evaluating images. The model relies on CLIP ViT-L/14 layers and combines logits from previous versions v3 and v4. This development is useful for researchers and developers who need to locally verify images without sending data to a server.

Key features

  • Works locally in the user's browser without sending pictures to hosting providers
  • Uses CLIP ViT-L/14 features from layers 10, 12, 14, 16, and 18 with three 224 px crops
  • Ensembles calibrated logits of versions v3 and v4 with weights of 0.25 and 0.75
  • Relies on the calibrated weight file head_v5.json and the clip_l14_b18_int8w_b32.onnx model

What you can do

  • Integrate the model into a web application for local verification of uploaded pictures in the browser
  • Set up an image classification pipeline using the ONNX runtime
  • Conduct research analysis of generation artifacts on the client side

Example

Use the model files clip_l14_b18_int8w_b32.onnx and head_v5.json in your ONNX-compatible runtime for feature extraction and final score calculation.

Read moreDetails
Backbone uses OpenAI CLIP via OpenCLIP (MIT)
Trained on OpenFake (CC BY-NC 4.0), MLLM-Generated-Image-Detection-Dataset, DOCCI, and other datasets
Non-commercial use (research preview)
Default ONNX model weight: 259,428,712 bytes
Discuss Telegram 1

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.2 GB · ready files0.8 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 GB
  • 4-bitweights 0.1 GB0.6 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # hashmap06/ai-image-detector-v5
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download hashmap06/ai-image-detector-v5 --local-dir ~/models/ai-image-detector-v5
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to ai-image-detector-v5
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of ai-image-detector-v5 (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

hashmap06/ai-image-detector-v5
Author
hashmap06
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

01

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments