Hugging Face's logo

qualcomm
/
Whisper-Small-Quantized

Automatic Speech Recognition
PyTorch
foundation
android
Model card Files
xet
 Community
Whisper-Small-Quantized
File size: 23,975 Bytes
0e21411
 
4491b69
0e21411
 
 
 
 
 
 
 
 
3858fbe
0e21411
 
 
3858fbe
4491b69
3858fbe
 
 
748970b
 
 
3858fbe
 
 
 
 
 
 
 
 
4491b69
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3858fbe
 
 
 
 
 
4491b69
3858fbe
 
 
 
 
 
4491b69
3858fbe
 
 
 
 
 
 
 
7da19b9
3858fbe
 
 
 
4491b69
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7da19b9
4491b69
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7da19b9
4491b69
 
 
 
 
 
 
 
 
 
 
 
 
 
0e21411
 
 
 
 
 
 
 
 
 
 
 
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
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
---
library_name: pytorch
license: apache-2.0
tags:
- foundation
- android
pipeline_tag: automatic-speech-recognition

---

![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/web-assets/model_demo.png)

# Whisper-Small-Quantized: Optimized for Qualcomm Devices

We have applied w8a16 quantization to significantly enhance performance and efficiency. HuggingFace Whisper-Small ASR (Automatic Speech Recognition) model is a state-of-the-art system designed for transcribing spoken language into written text. This model is based on the transformer architecture and has been optimized for edge inference by replacing Multi-Head Attention (MHA) with Single-Head Attention (SHA) and linear layers with convolutional (conv) layers. It exhibits robust performance in realistic, noisy environments, making it highly reliable for real-world applications. Specifically, it excels in long-form transcription, capable of accurately transcribing audio clips up to 30 seconds long. Time to the first token is the encoder's latency, while time to each additional token is decoder's latency, where we assume a max decoded length specified below.

This is based on the implementation of Whisper-Small-Quantized found [here](https://github.com/huggingface/transformers/tree/v4.42.3/src/transformers/models/whisper).
This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.61.0/src/qai_hub_models/models/whisper_small_quantized) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).

Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) to run these models on a hosted Qualcomm® device.

## Deploying Whisper-Small-Quantized on-device

This model is compatible with the Qualcomm Voice AI SDK. Download the SDK from the [Qualcomm Package Manager](https://qpm.qualcomm.com/#/main/tools/details/VoiceAI_ASR) to deploy this model on-device.
## Getting Started
There are two ways to deploy this model on your device:

### Option 1: Download Pre-Exported Models

Below are pre-exported model assets ready for deployment.

| Runtime | Precision | Chipset | SDK Versions | Download |
|---|---|---|---|---|
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X2 Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_snapdragon_x2_elite.zip)
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X Elite | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_snapdragon_x_elite.zip)
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_snapdragon_8gen3.zip)
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_qcs8275.zip)
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_qcs8550_proxy.zip)
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_qcm6690.zip)
| PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_qcs9075.zip)
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_snapdragon_7gen4.zip)
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_snapdragon_8_elite_for_galaxy.zip)
| PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-precompiled_qnn_onnx-w8a16-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X2 Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_snapdragon_x2_elite.zip)
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_snapdragon_x_elite.zip)
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_snapdragon_8gen3.zip)
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_qcs8275.zip)
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_qcs8550_proxy.zip)
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8775P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_sa8775p.zip)
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-6690 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_qcm6690.zip)
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_qcs9075.zip)
| QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA7255P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_sa7255p.zip)
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 7 Gen 4 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_snapdragon_7gen4.zip)
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_snapdragon_8_elite_for_galaxy.zip)
| QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-qnn_context_binary-w8a16-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
| VOICE_AI | w8a16 | Snapdragon® X2 Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_snapdragon_x2_elite.zip)
| VOICE_AI | w8a16 | Snapdragon® X Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_snapdragon_x_elite.zip)
| VOICE_AI | w8a16 | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_snapdragon_8gen3.zip)
| VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_qcs8275.zip)
| VOICE_AI | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_qcs8550_proxy.zip)
| VOICE_AI | w8a16 | Qualcomm® SA8775P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_sa8775p.zip)
| VOICE_AI | w8a16 | Qualcomm® Dragonwing™ Q-6690 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_qcm6690.zip)
| VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_qcs9075.zip)
| VOICE_AI | w8a16 | Qualcomm® SA7255P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_sa7255p.zip)
| VOICE_AI | w8a16 | Snapdragon® 7 Gen 4 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_snapdragon_7gen4.zip)
| VOICE_AI | w8a16 | Snapdragon® 8 Elite Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_snapdragon_8_elite_for_galaxy.zip)
| VOICE_AI | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/whisper_small_quantized/releases/v0.61.0/whisper_small_quantized-voice_ai-w8a16-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)

For more device-specific assets and performance metrics, visit **[Whisper-Small-Quantized on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/whisper_small_quantized)**.


### Option 2: Export with Custom Configurations

Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.61.0/src/qai_hub_models/models/whisper_small_quantized) Python library to compile and export the model with your own:
- Custom weights (e.g., fine-tuned checkpoints)
- Custom input shapes
- Target device and runtime configurations

This option is ideal if you need to customize the model beyond the default configuration provided here.

See our repository for [Whisper-Small-Quantized on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.61.0/src/qai_hub_models/models/whisper_small_quantized) for usage instructions.

## Model Details

**Model Type:** Model_use_case.speech_recognition

**Model Stats:**
- Input resolution: 80x3000 (30 seconds audio)
- Max decoded sequence length: 200 tokens
- Model checkpoint: openai/whisper-small

## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X2 Elite | 3.794 ms | 33 - 33 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X Elite | 7.356 ms | 186 - 186 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 6.176 ms | 35 - 46 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 10.39 ms | 37 - 51 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 9.092 ms | 27 - 60 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 7.972 ms | 0 - 192 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® QCS8450 | 10.39 ms | 37 - 51 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 8.889 ms | 24 - 57 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 7.356 ms | 186 - 186 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 42.123 ms | 40 - 51 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 10.523 ms | 28 - 38 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 4.663 ms | 26 - 34 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 4.663 ms | 26 - 34 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 3.938 ms | 23 - 35 MB | NPU
| decoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 10.523 ms | 28 - 38 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X2 Elite | 4.316 ms | 30 - 30 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X Elite | 7.543 ms | 30 - 30 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Gen 3 Mobile | 6.038 ms | 30 - 38 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 8.839 ms | 25 - 61 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 12.871 ms | 21 - 30 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 7.877 ms | 30 - 32 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8775P | 8.995 ms | 16 - 26 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8650P | 8.995 ms | 16 - 26 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8255P | 8.995 ms | 16 - 26 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 8.735 ms | 25 - 60 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 7.543 ms | 30 - 30 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 34.495 ms | 30 - 37 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 10.242 ms | 30 - 36 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 4.605 ms | 30 - 39 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA7255P | 12.871 ms | 21 - 30 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Mobile | 4.605 ms | 30 - 39 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 3.903 ms | 25 - 35 MB | NPU
| decoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 7 Gen 4 Mobile | 10.242 ms | 30 - 36 MB | NPU
| decoder | VOICE_AI | w8a16 | Snapdragon® X2 Elite | 4.358 ms | 30 - 30 MB | NPU
| decoder | VOICE_AI | w8a16 | Snapdragon® X Elite | 7.353 ms | 30 - 30 MB | NPU
| decoder | VOICE_AI | w8a16 | Snapdragon® 8 Gen 3 Mobile | 5.992 ms | 30 - 38 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 8.877 ms | 25 - 61 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 12.776 ms | 14 - 23 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 7.96 ms | 30 - 35 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® SA8775P | 8.942 ms | 19 - 29 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® SA8650P | 8.942 ms | 19 - 29 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® SA8255P | 8.942 ms | 19 - 29 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 10.043 ms | 27 - 62 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 7.353 ms | 30 - 30 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 34.793 ms | 20 - 27 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 10.328 ms | 28 - 34 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 4.6 ms | 30 - 43 MB | NPU
| decoder | VOICE_AI | w8a16 | Qualcomm® SA7255P | 12.776 ms | 14 - 23 MB | NPU
| decoder | VOICE_AI | w8a16 | Snapdragon® 8 Elite Mobile | 4.6 ms | 30 - 43 MB | NPU
| decoder | VOICE_AI | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 3.935 ms | 24 - 33 MB | NPU
| decoder | VOICE_AI | w8a16 | Snapdragon® 7 Gen 4 Mobile | 10.328 ms | 28 - 34 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X2 Elite | 155.09 ms | 67 - 67 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® X Elite | 300.906 ms | 127 - 127 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 270.196 ms | 63 - 75 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 361.383 ms | 55 - 69 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 305.09 ms | 62 - 66 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 394.726 ms | 0 - 130 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® QCS8450 | 361.383 ms | 55 - 69 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 295.827 ms | 63 - 66 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 300.906 ms | 127 - 127 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 4465.237 ms | 26 - 37 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 487.802 ms | 57 - 69 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 222.951 ms | 43 - 55 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 222.951 ms | 43 - 55 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 185.642 ms | 63 - 75 MB | NPU
| encoder | PRECOMPILED_QNN_ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 487.802 ms | 57 - 69 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X2 Elite | 157.203 ms | 0 - 0 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® X Elite | 299.672 ms | 0 - 0 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Gen 3 Mobile | 270.846 ms | 1 - 8 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 305.02 ms | 0 - 29 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 513.208 ms | 1 - 10 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 391.679 ms | 1 - 3 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8775P | 315.075 ms | 1 - 10 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8650P | 315.075 ms | 1 - 10 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA8255P | 315.075 ms | 1 - 10 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 295.566 ms | 0 - 29 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 299.672 ms | 0 - 0 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 4456.344 ms | 1 - 13 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 472.204 ms | 0 - 7 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 224.703 ms | 1 - 12 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Qualcomm® SA7255P | 513.208 ms | 1 - 10 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Mobile | 224.703 ms | 1 - 12 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 176.632 ms | 1 - 10 MB | NPU
| encoder | QNN_CONTEXT_BINARY | w8a16 | Snapdragon® 7 Gen 4 Mobile | 472.204 ms | 0 - 7 MB | NPU
| encoder | VOICE_AI | w8a16 | Snapdragon® X2 Elite | 155.611 ms | 0 - 0 MB | NPU
| encoder | VOICE_AI | w8a16 | Snapdragon® X Elite | 302.717 ms | 0 - 0 MB | NPU
| encoder | VOICE_AI | w8a16 | Snapdragon® 8 Gen 3 Mobile | 268.009 ms | 3 - 10 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 306.202 ms | 0 - 29 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 516.083 ms | 1 - 10 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 391.78 ms | 1 - 10 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® SA8775P | 314.484 ms | 1 - 10 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® SA8650P | 314.484 ms | 1 - 10 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® SA8255P | 314.484 ms | 1 - 10 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 295.173 ms | 0 - 29 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 302.717 ms | 0 - 0 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 3675.066 ms | 1 - 13 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 482.332 ms | 1 - 7 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 223.821 ms | 1 - 14 MB | NPU
| encoder | VOICE_AI | w8a16 | Qualcomm® SA7255P | 516.083 ms | 1 - 10 MB | NPU
| encoder | VOICE_AI | w8a16 | Snapdragon® 8 Elite Mobile | 223.821 ms | 1 - 14 MB | NPU
| encoder | VOICE_AI | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 174.948 ms | 1 - 11 MB | NPU
| encoder | VOICE_AI | w8a16 | Snapdragon® 7 Gen 4 Mobile | 482.332 ms | 1 - 7 MB | NPU

## License
* The license for the original implementation of Whisper-Small-Quantized can be found
  [here](https://github.com/huggingface/transformers/blob/v4.42.3/LICENSE).

## References
* [Robust Speech Recognition via Large-Scale Weak Supervision](https://cdn.openai.com/papers/whisper.pdf)
* [Source Model Implementation](https://github.com/huggingface/transformers/tree/v4.42.3/src/transformers/models/whisper)

## Community
* Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
* For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).