Skip to content
Audio

BAMM CNN detectors

CNN, которая ищет AI-музыку в телеэфире

2 min read

Два бинарных классификатора от DaveLoay оценивают вероятность того, что музыка сгенерирована AI (Suno v3.5), а не создана человеком. Это baseline-модели статьи ISMIR 2026 о детекции AI-музыки в реальном телевещании: 1,64 млн параметров, Keras, вес 6,6 МБ.

Ключевые особенности

  • CNN Clean обучен на чистой музыке: FMA-medium (человек) против SONICS Suno v3.5 (AI).
  • CNN Broadcast — та же музыка, но 70% каждого класса смешано с речью LibriSpeech при случайном SNR от −30 до +30 дБ, 30% оставлено чистыми; всё перекодировано в 8 кГц / 40 кбит/с MP3.
  • На выходе P(AI) — среднее по непересекающимся 5-секундным окнам; в статье порог AI — P(AI) >= 0.5.
  • Архитектура: шесть блоков Conv2D 3×3 (16–512 фильтров), max pooling, batch norm, global average pooling, dropout, dense 128 со swish и sigmoid.
  • В каждом варианте: model.h5, config_snapshot.yaml, training_history.json и model_summary.txt.
  • AI-класс — только Suno v3.5; на реальном эфире оба детектора заметно деградируют, особенно на фоновой музыке, и для мониторинга в одиночку их недостаточно.

Пример

git clone https://github.com/DaveLoay/BAMM-dataset.git && cd BAMM-dataset
pip install -e .
python scripts/download_weights.py
python -m pipelines.predict --model-path weights/cnn_broadcast/model.h5 --input clip.mp4
ПодробнееCFM: F1 0,992 и ROC-AUC 0,998 у CNN Clean; у CNN Broadcast — 0,992 и 0,996.
STB: CNN Broadcast F1 0,661, ROC-AUC 0,926; RTB по клипам — F1 0,472, ROC-AUC 0,796.
Нужны ffmpeg, TensorFlow 2.11–<2.16, librosa и numpy<2; модель видит только сигнал ниже 4 кГц.
Веса под CC BY-NC 4.0, код на GitHub — MIT. Не используйте одну оценку как единственное основание для решений о человеке или работе.
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.0 GB · ready files0.5 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 GB
  • 4-bitweights 0.0 GB0.5 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.

    # DaveLoay/bamm-cnn-detectors
    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 DaveLoay/bamm-cnn-detectors --local-dir ~/models/bamm-cnn-detectors
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to bamm-cnn-detectors
  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 bamm-cnn-detectors (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

DaveLoay/bamm-cnn-detectors
Author
DaveLoay
Category
Audio
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

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

Comments