CE

mispeech/ced-tiny

CedGGUF by mudlerUpdated 2026-06-21
Download on Hugging Face
Parameters86M
Context—
Layers—
Architectureced
Smallest quant0.0 GB
Quick answer

mispeech/ced-tiny needs about 0.5 GB of memory at Q8_0 (a 0.0 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 9 quants, from 0.0 GB (Q8_0) to 0.3 GB (F32).

CED‑Tiny: 85.7M‑parameter audio‑tagging model in GGUF format

CED‑Tiny is a 0.0857‑billion‑parameter audio‑tagging model based on the CED architecture. It supports 527 AudioSet classes and is available in multiple GGUF quantizations. The model is licensed under Apache‑2.0 and can run on CPUs via ced.cpp or LocalAI.

Good at

  • Fast inference on low‑end CPUs
  • Efficient 527‑class audio tagging
  • Small file size for edge deployment

Limits

  • No vision support
  • Context length limited by 8k tokens
  • Requires 8k context for memory estimates

Best for

  • Edge audio classification on Raspberry‑Pi
  • Real‑time sound event detection
  • Low‑memory inference on CPUs

Who it’s for. Developers building audio‑tagging applications on resource‑constrained devices. Researchers needing a lightweight baseline for AudioSet experiments.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q8_00.0 GB 0.5 GB
Fits Get file →
Q8_00.0 GB 0.5 GB
Fits Get file →
F160.0 GB 0.5 GB
Fits Get file →
F160.0 GB 0.5 GB
Fits Get file →
Q8_00.0 GB 0.5 GB
Fits Get file →
F160.0 GB 0.5 GB
Fits Get file →
Q8_00.1 GB 0.6 GB
Fits Get file →
F160.2 GB 0.7 GB
Fits Get file →
F320.3 GB 0.9 GB
Fits Get file →

For 8 GB VRAM, the Q8_0 quantization at 0.01 GB file size uses about 0.5 GB memory at 8k context. If more accuracy is needed, choose F16 at 0.02 GB, still under 1 GB memory. Larger models like F32 use 0.34 GB file size and 0.9 GB memory.

What GPU do you need for mispeech/ced-tiny?

These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is F32 (0.3 GB file, about 0.9 GB in use).

Which mispeech/ced-tiny quant should you choose?

Q8_0 (0.0 GB) is the middle option in this repository. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.

If memory is tight, the smallest file is Q8_0 at 0.0 GB, with the quality trade-offs that come with fewer bits. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureced
Parameters86M

Source

Quantized bymudler
FamilyCed
Repositorymudler/ced-gguf
Licenseapache-2.0
Downloads (HF)9,706
Last updated2026-06-21

Frequently asked questions about mispeech/ced-tiny

How much VRAM does mispeech/ced-tiny need?

About 0.5 GB at Q8_0 with an 8K context window. That is the 0.0 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run mispeech/ced-tiny on an 8 GB GPU?

Yes, with F32: the 0.3 GB file needs about 0.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of mispeech/ced-tiny GGUF?

It depends on the quant: Q8_0 is 0.0 GB, Q8_0 is 0.0 GB, F16 is 0.0 GB, F16 is 0.0 GB, Q8_0 is 0.0 GB, F16 is 0.0 GB and 3 more.

Where can I download mispeech/ced-tiny GGUF?

The files are on Hugging Face in the mudler/ced-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does mispeech/ced-tiny use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run CED‑Tiny on a Raspberry‑Pi with 1 GB RAM?

Yes, the Q8_0 0.01 GB file uses about 0.5 GB memory at 8k context, fitting within 1 GB RAM.

Does CED‑Tiny support real‑time audio streaming?

Yes, ced.cpp can stream over a websocket API for live recognition.

What license applies to the model weights?

The weights are released under Apache‑2.0.

Similar models

Browse all →

More from mudler

All by mudler →

More Ced models

All Ced →

Other tiny models

All tiny models →