File size: 946 Bytes
946748c 277918f 946748c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 | schema_version: 1.0.0
report_type: speech-asr
task_type: automatic-speech-recognition
title: Whisper Small INT8 — ExecuTorch + XNNPACK
description: >-
A compact INT8-quantized speech recognition model for transcribing 16 kHz audio, using an
encoder-decoder Transformer that converts log-mel spectrograms into text tokens. It combines dynamic
INT8 linear layers with selective FP32 components to reduce size and improve inference efficiency
while preserving transcription quality.
id: Arm/whisper-small-int8-xnnpack-executorch
filename: whisper-small-int8-executorch.pte
base_model_id: openai/whisper-small
profile: Arm-Optimized
weight_dtype: int8
quantization:
method: PTQ-dynamic
weight_bits: 8
activation_bits: 8
symmetric: true
mode: dynamic
weight_granularity: per-channel
variant: 8da8w + embed_tokens weight-only int8
fp32_layers:
- embed_positions
model_size_mb: 395.046
parameter_count: 241734912
format: pte
|