Skip to content
Audio

SUPlime-L Core ML

диаризация спикеров на Apple Silicon

1 min read

Конвертация моделей диаризации спикеров из rewayai/suplime-large в Core ML (mlprogram, fp16, скомпилированные .mlmodelc) для работы на устройстве в macOS 15+ с Apple Silicon. Автор — giuseppe-trisciuoglio; конвертация сделана для MeetRecord. Сегментация: WavLM-Large + Conformer, 349 млн параметров; эмбеддинги: SimAM-ResNet34, 25 млн параметров. Без дообучения.

Ключевые особенности

  • Два модуля: powerset-сегментация и эмбеддинги спикеров.
  • Фиксированный вход: 10 с, 16 кГц, моно; сегментация принимает [1, 1, 160000] и выдаёт [1, 499, 11].
  • Кластеризация agglomerative, centroid linkage, cosine distance, порог 0.72, min_cluster_size 12.
  • На M2 Pro сегментация на cpuAndGPU — около 0.12 с на окно, эмбеддинги на all — около 0.04 с.
  • Только некоммерческое использование: CC BY-NC 4.0.

Пример

Сценарий: подать 16 кГц mono float32 в диапазоне [-1, 1], нарезать окна по 160000 сэмплов с шагом 16000, усреднить перекрывающиеся кадры сегментации, посчитать kaldi fbank для чистых реплик, получить эмбеддинги и кластеризовать их по порогу 0.72.

ПодробнееСегментация: 707 МБ в fp16; fbank не входит в Core ML и его нужно реализовать отдельно.
На 9 синтетических 10-секундных окнах argmax совпал в 9/9; минимум cosine для эмбеддингов — 0.9998.
Итальянский не указан среди языков обучения; качество вне английского и китайского не задокументировано.
Лицензия требует атрибуции и указания изменений.
Discuss Telegram 0 37

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

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.

    # giuseppe-trisciuoglio/suplime-large-coreml
    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 giuseppe-trisciuoglio/suplime-large-coreml --local-dir ~/models/suplime-large-coreml
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to suplime-large-coreml
  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 suplime-large-coreml (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

giuseppe-trisciuoglio/suplime-large-coreml
Author
giuseppe-trisciuoglio
Category
Audio
Released
Posted

On NeyroLearn

Hugging Face

370

Ask AI about this model

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

Comments