Skip to content
Audio

Pocket TTS Core ML with timestamps: Speech synthesis with precise word-level timing

1 min readupdated

This model is a conversion of Kyutai's Pocket TTS speech engine into the Core ML format for running on Apple devices. The key feature of this version is the added export of word-level timestamps (the exact pronunciation time of each word) directly during speech generation, enabled by the attention weights of specific model heads. This is useful for mobile app and service developers who need synchronization between audio and text, such as for karaoke or real-time subtitle highlighting.

Key features

  • Optimized for Core ML to run on Apple devices
  • Added pre-softmax attention score output to obtain precise word timestamps
  • File structure compatible with the FluidInference ecosystem and FluidAudio library
  • Base weights and decoder components fully match the original model from Kyutai

What you can do

  • Integrate local speech synthesis into iOS or macOS apps with word-by-word subtitle display
  • Create an interactive player with text highlighting during audio playback
  • Use the model alongside the FluidAudio library for fast performance on Apple Silicon chips

Example

The model is used via compatible libraries like FluidAudio to generate speech from text while simultaneously obtaining word-level attention timestamps.

Read moreDetails
Base model: kyutai/pocket-tts
Library: coreml
License: CC-BY-4.0
Features: flowlm_step.mlmodelc re-exported with an additional timestamp output
Discuss Telegram 0 14

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.1 GB · ready files0.7 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 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.

    # tkpop777/pocket-tts-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 tkpop777/pocket-tts-coreml --local-dir ~/models/pocket-tts-coreml
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to pocket-tts-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 pocket-tts (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face

About the model

tkpop777/pocket-tts-coreml
Author
tkpop777
Category
Audio
Released
Posted

On NeyroLearn

Hugging Face

140

Ask AI about this model

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

Comments