Modelpublic

Arm/whisper-tiny-int8-xnnpack-executorch

sourceHugging Faceapache-2.0updated 15d agoView on Hugging Face

0likes52downloads

Model Card

Whisper Tiny optimized for Arm-based mobile CPUs with SME2

Whisper Tiny, an automatic speech recognition model, optimized with INT8 dynamic quantization and exported to ExecuTorch for inference on Arm-based mobile CPUs with SME2.

Summary

This repository contains an Arm-optimized version of openai/whisper-tiny for automatic speech recognition. The model is provided in ExecuTorch .pte format, targeting Mobile CPU systems.

This version is intended to demonstrate efficient inference on Arm-based platforms while preserving the original model's intended behavior. Arm has evaluated this model on librispeech_asr and measured performance on a representative evaluation target.

Key results

Area	Result
Model format	ExecuTorch `.pte`
Target device class	Mobile CPU
Reference device	vivo X300 (C1-Ultra, C1-Premium, C1-Pro, Android 16 / OriginOS 6)
Primary performance result	RTFx 5.99x
Accuracy result	WER 7.57%
Size / memory result	116.58 MB

Original model

Field	Value
Original model	openai/whisper-tiny
Original source	Hugging Face
Original developer	OpenAI
Original model card	openai/whisper-tiny
Original license	Apache-2.0

Model files

File	Description
`whisper_tiny_vivo_executorch_optimized.pte`	Arm-optimized model for deployment
`whisper_preprocessor.pte`	ExecuTorch module that computes the log-mel spectrogram from raw audio; loaded by `example.py` when present, with a Python-side fallback otherwise
`tokenizer.json`	Fast Whisper tokenizer vocabulary and tokenisation rules
`tokenizer_config.json`	Whisper tokenizer configuration
`special_tokens_map.json`	Whisper special-token definitions
`pte_original/`	Original baseline model and tokenizer artifacts; not used by the optimized example
`example.py`	Minimal inference example
`pyproject.toml`	Pinned runtime dependencies for `example.py`, resolved with uv
`uv.lock`	Locked dependency resolution for `pyproject.toml`
`config.yaml`	Model I/O contract used by the example
`benchmarks/`	FP32 baseline and Arm-optimized benchmark records

Performance

Performance was measured on the reference configuration below. Results are intended to make the optimization reproducible but do not guarantee identical performance on every Arm-based system.

Reference configuration

Field	Value
Device / platform	vivo X300
CPU / accelerator	C1-Ultra, C1-Premium, C1-Pro (8 cores), CPU execution backend
OS	android, Android 16 / OriginOS 6
Runtime	ExecuTorch 1.1.0
Backend / delegate	XNNPACK, KleidiAI
Batch size	1
Precision	INT8 (PTQ-dynamic) — per-channel symmetric weights, dynamically quantized symmetric activations (8da8w)
Runs	10 warmup + 50 measured

Performance results

Metric	Original / baseline	Arm-optimized	Improvement
p50 latency	1547.0 ms	1169.5 ms	1.32x faster
p90 latency	1571.0 ms	1200.0 ms	1.31x faster
p99 latency	1577.0 ms	1203.0 ms	1.31x faster
Model size	220.26 MB	116.58 MB	1.89x smaller
Peak memory	3987.16 MB	3940.71 MB	1.01x less
RTFx	4.52x	5.99x	1.32x

Accuracy

Accuracy was evaluated using the same preprocessing, input resolution, and evaluation protocol described below. Where possible, the optimized model is compared against the original model under the same evaluation conditions.

Evaluation setup

Field	Value
Dataset	librispeech_asr
Split	test-clean
Number of samples	2620
Metric(s)	WER, CER
Evaluation runtime	ExecuTorch

Accuracy results

Metric	Original / baseline	Arm-optimized	Change
WER	7.56%	7.57%	+0.01 pp
CER	3.11%	3.13%	+0.02 pp

Accuracy was measured using the evaluation setup described above. Users should re-evaluate the model on their own data before production use.

Arm optimization approach

Arm optimized this model for efficient inference on Arm-based platforms using a hardware-aware conversion and validation flow.

For this release, Arm used:

Optimization area	Applied?	Notes
Model conversion	Yes	Converted to ExecuTorch `.pte` via Optimum ExecuTorch
Quantization	Yes	PTQ-dynamic, 8-bit dynamic activations + 8-bit weights, symmetric, per-channel (TorchAO 8da8w)
Runtime/backend selection	Yes	Selected XNNPACK with KleidiAI acceleration
Graph/runtime compatibility updates	Yes	Performed as part of the ExecuTorch export pipeline (slice-scatter KV-cache rewrite required for export)
Accuracy validation	Yes	Compared against the original model or published baseline
Performance validation	Yes	Measured on the reference Arm platform

The goal of this process is to improve deployment characteristics such as latency, memory use, model size, and runtime compatibility while preserving the model's intended behavior. Detailed conversion scripts, calibration configuration, or backend-specific implementation details may be provided separately where appropriate.

Using this model

Install dependencies

Dependencies are declared in pyproject.toml, which ships with this repository. Resolve and install them into a local virtual environment with uv:

bash

uv python install
uv sync --frozen

Run the example

bash

uv run example.py

Note: The Python/uv example runs on AWS Graviton (Ubuntu arm64) to confirm runtime compatibility only, and is intended as a guideline for building an equivalent run script on Mobile CPU systems.

Expected input

Property	Value
Input shape	[1, 80, 3000]
Input type	float32
Input range	N/A (log-mel spectrogram features, not a normalized pixel range)
Preprocessing	Resample to 16 kHz, then extract an 80-bin log-mel spectrogram (25 ms window, 10 ms hop, 30 s duration)

Expected output

Property	Value
Output shape	variable
Output type	token IDs (int)
Postprocessing	Greedy decode up to 128 new tokens, then Whisper tokenizer decode with skipspecialtokens=True (language=en, task=transcribe)

Intended use

This model is intended for developers evaluating automatic speech recognition workloads on Arm-based platforms. It is suitable as a reference implementation for benchmarking, prototyping, and integration exploration.

Limitations

—Performance depends on the target device, runtime version, backend/delegate support, memory configuration, and system load.
—Accuracy was evaluated on librispeech_asr test-clean and may not generalize to all domains.
—This release preserves the original model's intended task and behavior, but users should validate it for their own application, data, and deployment environment.
—This repository is not a replacement for the original model documentation.
—The log-mel spectrogram input is always padded or clipped to a fixed 30-second window; longer audio must be chunked externally before inference.

Additional notes

—Sample input: sample_input.flac is utterance 5338-24615-0014 of the LibriSpeech ASR corpus (dev-clean split) by Panayotov et al., via OpenSLR (CC BY 4.0).

About this version

Original Model: openai/whisper-tiny by OpenAI - Repository

Optimization/conversion: Arm-Optimized version for execution on Arm-based platforms.

Converted/optimized by: Arm

License: The Original Model and the Optimized Model are subject to Apache-2.0.

This repository contains a converted or optimized version of the Original Model (the “Optimized Model”). The Original Model has been converted or optimized as described above for execution on Arm-based platforms.

No retraining or fine-tuning of the Original Model was performed as part of the conversion or optimization. The conversion or optimization was not intended to change the Original Model’s behavior or intended use.

Original Model and Documentation

For information about the Original Model, including its development, training data, intended uses, limitations and other relevant information, please refer to the Original Model repository. Information in that repository was provided by the original developer or other third parties and, unless expressly stated otherwise, has not been independently verified by Arm.

Licenses and Third-Party Terms

Use of the Original Model and the Optimized Model is subject to the applicable licenses, usage restrictions and other terms identified above and in the relevant repositories. Publication of the Optimized Model does not grant any rights beyond those provided under the applicable license terms.

You are responsible for reviewing those terms and ensuring that your use of the Original Model and the Optimized Model is permitted.

Purpose of this Release

The Optimized Model is provided as a reference implementation to demonstrate and evaluate execution and performance on Arm-based systems. It is not a production-ready or supported solution.

Arm’s publication of the Optimized Model does not constitute an endorsement or certification of the Original Model or a representation that the Optimized Model is suitable for production use or any particular purpose.

To the fullest extent permitted by applicable law (i) the Optimized Model is provided “as is.” Arm makes no representations or warranties that the Original Model, the Optimized Model or their outputs are accurate, safe, secure, non-infringing, legally compliant, suitable for production use or fit for any particular purpose; and (ii) Arm will not be liable for any loss or damage arising from or in connection with the Optimized Model, its use or its outputs.

You are responsible for independently evaluating the Optimized Model, its outputs and its suitability for your intended use, including compliance with applicable legal, regulatory, safety and security requirements.

Arm does not commit to provide ongoing support, maintenance or updates for the Optimized Model. Any use of or reliance on the Optimized Model or its outputs is at your own risk.