qaihm-bot commited on
Commit
43cb0a9
·
verified ·
1 Parent(s): fc3a287

See https://github.com/qualcomm/ai-hub-models/releases/v0.63.0 for changelog.

Files changed (2) hide show
  1. README.md +174 -102
  2. release_assets.json +183 -43
README.md CHANGED
@@ -15,7 +15,7 @@ pipeline_tag: automatic-speech-recognition
15
  Zipformer streaming ASR (Automatic Speech Recognition) model is a state-of-the-art system designed for transcribing spoken language into written text streamingly. This model is based on the transformer architecture and has been optimized for edge inference by replacing 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 audios. Time to the first token is the encoder's latency, while time to each additional token is joiner's latency, where we assume a max decoded length specified below.
16
 
17
  This is based on the implementation of Zipformer found [here](https://github.com/k2-fsa/icefall).
18
- 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/v0.62.2/src/qai_hub_models/models/zipformer) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
19
 
20
  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.
21
 
@@ -28,41 +28,59 @@ Below are pre-exported model assets ready for deployment.
28
 
29
  | Runtime | Precision | Chipset | SDK Versions | Download |
30
  |---|---|---|---|---|
31
- | PRECOMPILED_QNN_ONNX | float | 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/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_x2_elite.zip)
32
- | PRECOMPILED_QNN_ONNX | float | 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/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_x_elite.zip)
33
- | PRECOMPILED_QNN_ONNX | float | 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/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen3.zip)
34
- | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.45, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen1.zip)
35
- | PRECOMPILED_QNN_ONNX | float | 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/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs8275.zip)
36
- | PRECOMPILED_QNN_ONNX | float | 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/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs8550_proxy.zip)
37
- | PRECOMPILED_QNN_ONNX | float | 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/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs9075.zip)
38
- | PRECOMPILED_QNN_ONNX | float | 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/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
39
- | PRECOMPILED_QNN_ONNX | float | 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/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
40
- | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_x2_elite.zip)
41
- | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_x_elite.zip)
42
- | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8gen3.zip)
43
- | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8gen1.zip)
44
- | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_qcs8275.zip)
45
- | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_qcs8550_proxy.zip)
46
- | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_sa8775p.zip)
47
- | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_qcs9075.zip)
48
- | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_sa7255p.zip)
49
- | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_sa8295p.zip)
50
- | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
51
- | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | QAIRT 2.45 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
52
 
53
  For more device-specific assets and performance metrics, visit **[Zipformer on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/zipformer)**.
54
 
55
 
56
  ### Option 2: Export with Custom Configurations
57
 
58
- Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.62.2/src/qai_hub_models/models/zipformer) Python library to compile and export the model with your own:
59
  - Custom weights (e.g., fine-tuned checkpoints)
60
  - Custom input shapes
61
  - Target device and runtime configurations
62
 
63
  This option is ideal if you need to customize the model beyond the default configuration provided here.
64
 
65
- See our repository for [Zipformer on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.62.2/src/qai_hub_models/models/zipformer) for usage instructions.
66
 
67
  ## Model Details
68
 
@@ -82,93 +100,147 @@ See our repository for [Zipformer on GitHub](https://github.com/qualcomm/ai-hub-
82
  ## Performance Summary
83
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
84
  |---|---|---|---|---|---|---
85
- | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 0.092 ms | 1 - 1 MB | NPU
86
- | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 0.142 ms | 7 - 7 MB | NPU
87
- | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.158 ms | 0 - 7 MB | NPU
88
- | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | 0.25 ms | 0 - 14 MB | NPU
89
- | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 0.357 ms | 0 - 4 MB | NPU
90
- | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.191 ms | 0 - 8 MB | NPU
91
- | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8450 | 0.25 ms | 0 - 14 MB | NPU
92
- | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 0.274 ms | 0 - 3 MB | NPU
93
- | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.142 ms | 7 - 7 MB | NPU
94
- | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 0.139 ms | 0 - 10 MB | NPU
95
- | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Mobile | 0.139 ms | 0 - 10 MB | NPU
96
- | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.137 ms | 0 - 7 MB | NPU
97
- | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 0.176 ms | 0 - 0 MB | NPU
98
- | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 0.241 ms | 0 - 0 MB | NPU
99
- | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 0.15 ms | 15 - 22 MB | NPU
100
- | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | 0.24 ms | 0 - 9 MB | NPU
101
- | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | 0.334 ms | 0 - 2 MB | NPU
102
- | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.197 ms | 0 - 1 MB | NPU
103
- | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | 0.324 ms | 0 - 10 MB | NPU
 
 
 
 
 
 
 
 
 
 
 
104
  | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8650P | 0.324 ms | 0 - 10 MB | NPU
105
  | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8255P | 0.324 ms | 0 - 10 MB | NPU
106
- | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 | 0.24 ms | 0 - 9 MB | NPU
107
- | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | 0.24 ms | 0 - 2 MB | NPU
108
- | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.241 ms | 0 - 0 MB | NPU
109
- | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ Q-8750 | 0.133 ms | 0 - 12 MB | NPU
110
- | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | 0.432 ms | 0 - 8 MB | NPU
111
- | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 0.456 ms | 0 - 5 MB | NPU
112
- | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | 0.133 ms | 0 - 12 MB | NPU
113
- | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.131 ms | 0 - 8 MB | NPU
114
- | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 7.921 ms | 10 - 10 MB | NPU
115
- | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 13.362 ms | 142 - 142 MB | NPU
116
- | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 8.885 ms | 9 - 21 MB | NPU
117
- | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | 16.148 ms | 9 - 23 MB | NPU
118
- | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 13.723 ms | 8 - 13 MB | NPU
119
- | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 12.508 ms | 0 - 159 MB | NPU
120
- | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8450 | 16.148 ms | 9 - 23 MB | NPU
121
- | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 14.85 ms | 8 - 12 MB | NPU
122
- | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 13.362 ms | 142 - 142 MB | NPU
123
- | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 7.533 ms | 9 - 22 MB | NPU
124
- | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Mobile | 7.533 ms | 9 - 22 MB | NPU
125
- | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 6.975 ms | 10 - 16 MB | NPU
126
- | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 8.417 ms | 3 - 3 MB | NPU
127
- | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 13.06 ms | 3 - 3 MB | NPU
128
- | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 8.737 ms | 2 - 12 MB | NPU
129
- | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | 16.003 ms | 2 - 13 MB | NPU
130
- | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | 13.446 ms | 3 - 8 MB | NPU
131
- | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 12.442 ms | 3 - 4 MB | NPU
132
- | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | 14.123 ms | 3 - 12 MB | NPU
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
133
  | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8650P | 14.123 ms | 3 - 12 MB | NPU
134
  | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8255P | 14.123 ms | 3 - 12 MB | NPU
135
- | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 | 16.003 ms | 2 - 13 MB | NPU
136
- | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | 14.075 ms | 3 - 7 MB | NPU
137
- | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-X7181 | 13.06 ms | 3 - 3 MB | NPU
138
- | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ Q-8750 | 7.486 ms | 3 - 15 MB | NPU
139
- | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | 28.792 ms | 3 - 11 MB | NPU
140
- | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 17.343 ms | 0 - 5 MB | NPU
141
- | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | 7.486 ms | 3 - 15 MB | NPU
142
- | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | 6.928 ms | 3 - 11 MB | NPU
 
 
 
 
 
 
 
 
 
143
  | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 0.08 ms | 1 - 1 MB | NPU
144
- | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 0.144 ms | 6 - 6 MB | NPU
145
- | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.155 ms | 0 - 6 MB | NPU
146
- | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | 0.238 ms | 0 - 14 MB | NPU
147
- | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 0.269 ms | 0 - 3 MB | NPU
148
- | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.188 ms | 0 - 8 MB | NPU
149
- | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8450 | 0.238 ms | 0 - 14 MB | NPU
150
- | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 0.27 ms | 0 - 3 MB | NPU
151
- | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.144 ms | 6 - 6 MB | NPU
152
- | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 0.134 ms | 0 - 7 MB | NPU
153
- | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Mobile | 0.134 ms | 0 - 7 MB | NPU
154
- | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.135 ms | 0 - 7 MB | NPU
155
- | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 0.204 ms | 0 - 0 MB | NPU
156
- | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 0.238 ms | 1 - 1 MB | NPU
157
- | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 0.149 ms | 0 - 7 MB | NPU
 
 
 
 
 
 
 
 
 
 
158
  | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | 0.229 ms | 0 - 9 MB | NPU
159
- | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | 0.242 ms | 0 - 2 MB | NPU
160
- | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.187 ms | 0 - 1 MB | NPU
161
- | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | 0.327 ms | 0 - 9 MB | NPU
162
  | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® SA8650P | 0.327 ms | 0 - 9 MB | NPU
163
  | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® SA8255P | 0.327 ms | 0 - 9 MB | NPU
164
  | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 | 0.229 ms | 0 - 9 MB | NPU
165
- | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | 0.234 ms | 2 - 4 MB | NPU
166
- | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.238 ms | 1 - 1 MB | NPU
167
- | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ Q-8750 | 0.134 ms | 0 - 13 MB | NPU
168
- | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | 0.475 ms | 0 - 8 MB | NPU
169
- | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 0.461 ms | 0 - 5 MB | NPU
170
- | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Mobile | 0.134 ms | 0 - 13 MB | NPU
171
- | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 Mobile | 0.134 ms | 0 - 8 MB | NPU
 
 
 
 
 
 
 
172
 
173
  ## License
174
  * The license for the original implementation of Zipformer can be found
 
15
  Zipformer streaming ASR (Automatic Speech Recognition) model is a state-of-the-art system designed for transcribing spoken language into written text streamingly. This model is based on the transformer architecture and has been optimized for edge inference by replacing 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 audios. Time to the first token is the encoder's latency, while time to each additional token is joiner's latency, where we assume a max decoded length specified below.
16
 
17
  This is based on the implementation of Zipformer found [here](https://github.com/k2-fsa/icefall).
18
+ 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/v0.63.0/src/qai_hub_models/models/zipformer) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
19
 
20
  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.
21
 
 
28
 
29
  | Runtime | Precision | Chipset | SDK Versions | Download |
30
  |---|---|---|---|---|
31
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
32
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
33
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_x2_elite.zip)
34
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_x_elite.zip)
35
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen3.zip)
36
+ | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen1.zip)
37
+ | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs8275.zip)
38
+ | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs8550_proxy.zip)
39
+ | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs9075.zip)
40
+ | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
41
+ | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_8_elite_for_galaxy.zip)
42
+ | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® X2 Elite | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_x2_elite.zip)
43
+ | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® X Elite | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_x_elite.zip)
44
+ | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_8gen3.zip)
45
+ | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_qcs8275.zip)
46
+ | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_qcs8550_proxy.zip)
47
+ | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.50, ONNX Runtime 1.27.1 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_qcs9075.zip)
48
+ | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
49
+ | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8_elite_for_galaxy.zip)
50
+ | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_x2_elite.zip)
51
+ | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_x_elite.zip)
52
+ | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8gen3.zip)
53
+ | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8gen1.zip)
54
+ | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_qcs8275.zip)
55
+ | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_qcs8550_proxy.zip)
56
+ | QNN_CONTEXT_BINARY | float | Qualcomm® SA8775P | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_sa8775p.zip)
57
+ | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_qcs9075.zip)
58
+ | QNN_CONTEXT_BINARY | float | Qualcomm® SA7255P | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_sa7255p.zip)
59
+ | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_sa8295p.zip)
60
+ | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip)
61
+ | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Elite For Galaxy Mobile | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_8_elite_for_galaxy.zip)
62
+ | QNN_CONTEXT_BINARY | mixed | Snapdragon® X2 Elite | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_x2_elite.zip)
63
+ | QNN_CONTEXT_BINARY | mixed | Snapdragon® X Elite | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_x_elite.zip)
64
+ | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Gen 3 Mobile | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_8gen3.zip)
65
+ | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-8275 | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_qcs8275.zip)
66
+ | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ QCS8550 (Proxy) | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_qcs8550_proxy.zip)
67
+ | QNN_CONTEXT_BINARY | mixed | Qualcomm® SA8775P | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_sa8775p.zip)
68
+ | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-9075 | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_qcs9075.zip)
69
+ | QNN_CONTEXT_BINARY | mixed | Qualcomm® SA7255P | QAIRT 2.50 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_sa7255p.zip)
70
 
71
  For more device-specific assets and performance metrics, visit **[Zipformer on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/zipformer)**.
72
 
73
 
74
  ### Option 2: Export with Custom Configurations
75
 
76
+ Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/v0.63.0/src/qai_hub_models/models/zipformer) Python library to compile and export the model with your own:
77
  - Custom weights (e.g., fine-tuned checkpoints)
78
  - Custom input shapes
79
  - Target device and runtime configurations
80
 
81
  This option is ideal if you need to customize the model beyond the default configuration provided here.
82
 
83
+ See our repository for [Zipformer on GitHub](https://github.com/qualcomm/ai-hub-models/blob/v0.63.0/src/qai_hub_models/models/zipformer) for usage instructions.
84
 
85
  ## Model Details
86
 
 
100
  ## Performance Summary
101
  | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
102
  |---|---|---|---|---|---|---
103
+ | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.14 ms | 0 - 7 MB | NPU
104
+ | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.139 ms | 0 - 7 MB | NPU
105
+ | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 0.108 ms | 1 - 1 MB | NPU
106
+ | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 0.144 ms | 7 - 7 MB | NPU
107
+ | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.153 ms | 0 - 7 MB | NPU
108
+ | decoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | 0.24 ms | 0 - 14 MB | NPU
109
+ | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 0.268 ms | 0 - 4 MB | NPU
110
+ | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.196 ms | 0 - 8 MB | NPU
111
+ | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8450 | 0.24 ms | 0 - 14 MB | NPU
112
+ | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 0.271 ms | 0 - 3 MB | NPU
113
+ | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.144 ms | 7 - 7 MB | NPU
114
+ | decoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 0.139 ms | 0 - 7 MB | NPU
115
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.135 ms | 0 - 7 MB | NPU
116
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Elite For Galaxy Mobile | 0.144 ms | 0 - 11 MB | NPU
117
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® X2 Elite | 0.083 ms | 1 - 1 MB | NPU
118
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® X Elite | 0.147 ms | 7 - 7 MB | NPU
119
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Gen 3 Mobile | 0.161 ms | 0 - 8 MB | NPU
120
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-8275 | 0.271 ms | 0 - 4 MB | NPU
121
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.195 ms | 0 - 8 MB | NPU
122
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-9075 | 0.307 ms | 0 - 3 MB | NPU
123
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-X7181 | 0.147 ms | 7 - 7 MB | NPU
124
+ | decoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ Q-8750 | 0.144 ms | 0 - 11 MB | NPU
125
+ | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.137 ms | 0 - 8 MB | NPU
126
+ | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.149 ms | 0 - 12 MB | NPU
127
+ | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 0.18 ms | 0 - 0 MB | NPU
128
+ | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 0.258 ms | 1 - 1 MB | NPU
129
+ | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 0.159 ms | 0 - 7 MB | NPU
130
+ | decoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | 0.237 ms | 0 - 9 MB | NPU
131
+ | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | 0.244 ms | 0 - 3 MB | NPU
132
+ | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.195 ms | 0 - 1 MB | NPU
133
  | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8650P | 0.324 ms | 0 - 10 MB | NPU
134
  | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8255P | 0.324 ms | 0 - 10 MB | NPU
135
+ | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 | 0.237 ms | 0 - 9 MB | NPU
136
+ | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | 0.242 ms | 0 - 2 MB | NPU
137
+ | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.258 ms | 1 - 1 MB | NPU
138
+ | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ Q-8750 | 0.149 ms | 0 - 12 MB | NPU
139
+ | decoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 0.362 ms | 0 - 6 MB | NPU
140
+ | decoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.131 ms | 0 - 8 MB | NPU
141
+ | decoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Elite For Galaxy Mobile | 0.135 ms | 0 - 12 MB | NPU
142
+ | decoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® X2 Elite | 0.176 ms | 0 - 0 MB | NPU
143
+ | decoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® X Elite | 0.255 ms | 1 - 1 MB | NPU
144
+ | decoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Gen 3 Mobile | 0.147 ms | 0 - 7 MB | NPU
145
+ | decoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-8275 | 0.333 ms | 0 - 3 MB | NPU
146
+ | decoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.194 ms | 0 - 1 MB | NPU
147
+ | decoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-9075 | 0.243 ms | 2 - 4 MB | NPU
148
+ | decoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-X7181 | 0.255 ms | 1 - 1 MB | NPU
149
+ | decoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ Q-8750 | 0.135 ms | 0 - 12 MB | NPU
150
+ | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 6.924 ms | 10 - 17 MB | NPU
151
+ | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 7.295 ms | 9 - 21 MB | NPU
152
+ | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 7.676 ms | 9 - 9 MB | NPU
153
+ | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 12.775 ms | 143 - 143 MB | NPU
154
+ | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 8.47 ms | 9 - 22 MB | NPU
155
+ | encoder | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | 15.306 ms | 9 - 24 MB | NPU
156
+ | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 13.361 ms | 8 - 14 MB | NPU
157
+ | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 12.016 ms | 9 - 11 MB | NPU
158
+ | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8450 | 15.306 ms | 9 - 24 MB | NPU
159
+ | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 14.025 ms | 8 - 13 MB | NPU
160
+ | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 12.775 ms | 143 - 143 MB | NPU
161
+ | encoder | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 7.295 ms | 9 - 21 MB | NPU
162
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 8.8 ms | 10 - 17 MB | NPU
163
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Elite For Galaxy Mobile | 9.628 ms | 9 - 20 MB | NPU
164
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® X2 Elite | 9.654 ms | 7 - 7 MB | NPU
165
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® X Elite | 17.646 ms | 75 - 75 MB | NPU
166
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Gen 3 Mobile | 13.461 ms | 9 - 17 MB | NPU
167
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-8275 | 18.008 ms | 8 - 14 MB | NPU
168
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 17.564 ms | 0 - 79 MB | NPU
169
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-9075 | 19.018 ms | 8 - 13 MB | NPU
170
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-X7181 | 17.646 ms | 75 - 75 MB | NPU
171
+ | encoder | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ Q-8750 | 9.628 ms | 9 - 20 MB | NPU
172
+ | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 6.901 ms | 3 - 10 MB | NPU
173
+ | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite For Galaxy Mobile | 7.28 ms | 3 - 11 MB | NPU
174
+ | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 8.2 ms | 3 - 3 MB | NPU
175
+ | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 12.744 ms | 3 - 3 MB | NPU
176
+ | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 8.25 ms | 2 - 11 MB | NPU
177
+ | encoder | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | 14.672 ms | 3 - 13 MB | NPU
178
+ | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | 13.239 ms | 3 - 8 MB | NPU
179
+ | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 11.9 ms | 3 - 5 MB | NPU
180
  | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8650P | 14.123 ms | 3 - 12 MB | NPU
181
  | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8255P | 14.123 ms | 3 - 12 MB | NPU
182
+ | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 | 14.672 ms | 3 - 13 MB | NPU
183
+ | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | 14.261 ms | 3 - 8 MB | NPU
184
+ | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-X7181 | 12.744 ms | 3 - 3 MB | NPU
185
+ | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ Q-8750 | 7.28 ms | 3 - 11 MB | NPU
186
+ | encoder | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 16.987 ms | 0 - 6 MB | NPU
187
+ | encoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 8.816 ms | 3 - 12 MB | NPU
188
+ | encoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Elite For Galaxy Mobile | 9.63 ms | 3 - 14 MB | NPU
189
+ | encoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® X2 Elite | 10.336 ms | 3 - 3 MB | NPU
190
+ | encoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® X Elite | 17.364 ms | 3 - 3 MB | NPU
191
+ | encoder | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Gen 3 Mobile | 13.077 ms | 2 - 10 MB | NPU
192
+ | encoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-8275 | 17.876 ms | 3 - 9 MB | NPU
193
+ | encoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 17.483 ms | 3 - 5 MB | NPU
194
+ | encoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-9075 | 19.364 ms | 5 - 10 MB | NPU
195
+ | encoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-X7181 | 17.364 ms | 3 - 3 MB | NPU
196
+ | encoder | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ Q-8750 | 9.63 ms | 3 - 14 MB | NPU
197
+ | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.137 ms | 0 - 7 MB | NPU
198
+ | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.137 ms | 0 - 11 MB | NPU
199
  | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® X2 Elite | 0.08 ms | 1 - 1 MB | NPU
200
+ | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® X Elite | 0.147 ms | 6 - 6 MB | NPU
201
+ | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 3 Mobile | 0.16 ms | 0 - 8 MB | NPU
202
+ | joiner | PRECOMPILED_QNN_ONNX | float | Snapdragon® 8 Gen 1 Mobile | 0.239 ms | 0 - 14 MB | NPU
203
+ | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 0.349 ms | 0 - 4 MB | NPU
204
+ | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.195 ms | 0 - 8 MB | NPU
205
+ | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® QCS8450 | 0.239 ms | 0 - 14 MB | NPU
206
+ | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 0.269 ms | 0 - 3 MB | NPU
207
+ | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.147 ms | 6 - 6 MB | NPU
208
+ | joiner | PRECOMPILED_QNN_ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 0.137 ms | 0 - 11 MB | NPU
209
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.138 ms | 0 - 7 MB | NPU
210
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Elite For Galaxy Mobile | 0.142 ms | 0 - 8 MB | NPU
211
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® X2 Elite | 0.085 ms | 1 - 1 MB | NPU
212
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® X Elite | 0.15 ms | 7 - 7 MB | NPU
213
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Snapdragon® 8 Gen 3 Mobile | 0.155 ms | 0 - 8 MB | NPU
214
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-8275 | 0.301 ms | 0 - 4 MB | NPU
215
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.197 ms | 0 - 8 MB | NPU
216
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-9075 | 0.305 ms | 0 - 3 MB | NPU
217
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ IQ-X7181 | 0.15 ms | 7 - 7 MB | NPU
218
+ | joiner | PRECOMPILED_QNN_ONNX | mixed | Qualcomm® Dragonwing™ Q-8750 | 0.142 ms | 0 - 8 MB | NPU
219
+ | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.132 ms | 0 - 7 MB | NPU
220
+ | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Elite For Galaxy Mobile | 0.132 ms | 0 - 12 MB | NPU
221
+ | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® X2 Elite | 0.212 ms | 0 - 0 MB | NPU
222
+ | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® X Elite | 0.239 ms | 1 - 1 MB | NPU
223
+ | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 3 Mobile | 0.154 ms | 0 - 8 MB | NPU
224
  | joiner | QNN_CONTEXT_BINARY | float | Snapdragon® 8 Gen 1 Mobile | 0.229 ms | 0 - 9 MB | NPU
225
+ | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-8275 | 0.291 ms | 0 - 3 MB | NPU
226
+ | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.193 ms | 0 - 1 MB | NPU
 
227
  | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® SA8650P | 0.327 ms | 0 - 9 MB | NPU
228
  | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® SA8255P | 0.327 ms | 0 - 9 MB | NPU
229
  | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® QCS8450 | 0.229 ms | 0 - 9 MB | NPU
230
+ | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-9075 | 0.239 ms | 0 - 2 MB | NPU
231
+ | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ IQ-X7181 | 0.239 ms | 1 - 1 MB | NPU
232
+ | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® Dragonwing™ Q-8750 | 0.132 ms | 0 - 12 MB | NPU
233
+ | joiner | QNN_CONTEXT_BINARY | float | Qualcomm® SA8295P | 0.346 ms | 0 - 5 MB | NPU
234
+ | joiner | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 0.134 ms | 0 - 8 MB | NPU
235
+ | joiner | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Elite For Galaxy Mobile | 0.142 ms | 0 - 12 MB | NPU
236
+ | joiner | QNN_CONTEXT_BINARY | mixed | Snapdragon® X2 Elite | 0.189 ms | 0 - 0 MB | NPU
237
+ | joiner | QNN_CONTEXT_BINARY | mixed | Snapdragon® X Elite | 0.256 ms | 0 - 0 MB | NPU
238
+ | joiner | QNN_CONTEXT_BINARY | mixed | Snapdragon® 8 Gen 3 Mobile | 0.162 ms | 0 - 8 MB | NPU
239
+ | joiner | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-8275 | 0.245 ms | 0 - 3 MB | NPU
240
+ | joiner | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 0.194 ms | 0 - 1 MB | NPU
241
+ | joiner | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-9075 | 0.243 ms | 0 - 2 MB | NPU
242
+ | joiner | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ IQ-X7181 | 0.256 ms | 0 - 0 MB | NPU
243
+ | joiner | QNN_CONTEXT_BINARY | mixed | Qualcomm® Dragonwing™ Q-8750 | 0.142 ms | 0 - 12 MB | NPU
244
 
245
  ## License
246
  * The license for the original implementation of Zipformer can be found
release_assets.json CHANGED
@@ -1,165 +1,305 @@
1
  {
2
- "version": "0.62.2",
3
  "precisions": {
4
  "float": {
5
  "chipset_assets": {
6
  "qualcomm-qcs8275": {
7
  "precompiled_qnn_onnx": {
8
  "tool_versions": {
9
- "qairt": "2.45.0.260326154327",
10
  "onnx_runtime": "1.27.1"
11
  },
12
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs8275.zip"
13
  },
14
  "qnn_context_binary": {
15
  "tool_versions": {
16
- "qairt": "2.45.0.260326154327"
17
  },
18
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_qcs8275.zip"
19
  }
20
  },
21
  "qualcomm-qcs8550-proxy": {
22
  "precompiled_qnn_onnx": {
23
  "tool_versions": {
24
- "qairt": "2.45.0.260326154327",
25
  "onnx_runtime": "1.27.1"
26
  },
27
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs8550_proxy.zip"
28
  },
29
  "qnn_context_binary": {
30
  "tool_versions": {
31
- "qairt": "2.45.0.260326154327"
32
  },
33
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_qcs8550_proxy.zip"
34
  }
35
  },
36
  "qualcomm-qcs9075": {
37
  "precompiled_qnn_onnx": {
38
  "tool_versions": {
39
- "qairt": "2.45.0.260326154327",
40
  "onnx_runtime": "1.27.1"
41
  },
42
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs9075.zip"
43
  },
44
  "qnn_context_binary": {
45
  "tool_versions": {
46
- "qairt": "2.45.0.260326154327"
47
  },
48
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_qcs9075.zip"
49
  }
50
  },
51
  "qualcomm-sa7255p": {
52
  "qnn_context_binary": {
53
  "tool_versions": {
54
- "qairt": "2.45.0.260326154327"
55
  },
56
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_sa7255p.zip"
57
  }
58
  },
59
  "qualcomm-sa8295p": {
60
  "qnn_context_binary": {
61
  "tool_versions": {
62
- "qairt": "2.45.0.260326154327"
63
  },
64
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_sa8295p.zip"
65
  }
66
  },
67
  "qualcomm-sa8775p": {
68
  "qnn_context_binary": {
69
  "tool_versions": {
70
- "qairt": "2.45.0.260326154327"
71
  },
72
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_sa8775p.zip"
73
  }
74
  },
75
  "qualcomm-snapdragon-8-elite-for-galaxy": {
76
  "precompiled_qnn_onnx": {
77
  "tool_versions": {
78
- "qairt": "2.45.0.260326154327",
79
  "onnx_runtime": "1.27.1"
80
  },
81
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_for_galaxy.zip"
82
  },
83
  "qnn_context_binary": {
84
  "tool_versions": {
85
- "qairt": "2.45.0.260326154327"
86
  },
87
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8_elite_for_galaxy.zip"
88
  }
89
  },
90
  "qualcomm-snapdragon-8-elite-gen5-for-galaxy": {
91
  "precompiled_qnn_onnx": {
92
  "tool_versions": {
93
- "qairt": "2.45.0.260326154327",
94
  "onnx_runtime": "1.27.1"
95
  },
96
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
97
  },
98
  "qnn_context_binary": {
99
  "tool_versions": {
100
- "qairt": "2.45.0.260326154327"
101
  },
102
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
103
  }
104
  },
105
  "qualcomm-snapdragon-8gen1": {
106
  "precompiled_qnn_onnx": {
107
  "tool_versions": {
108
- "qairt": "2.45.0.260326154327",
109
  "onnx_runtime": "1.27.1"
110
  },
111
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen1.zip"
112
  },
113
  "qnn_context_binary": {
114
  "tool_versions": {
115
- "qairt": "2.45.0.260326154327"
116
  },
117
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8gen1.zip"
118
  }
119
  },
120
  "qualcomm-snapdragon-8gen3": {
121
  "precompiled_qnn_onnx": {
122
  "tool_versions": {
123
- "qairt": "2.45.0.260326154327",
124
  "onnx_runtime": "1.27.1"
125
  },
126
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen3.zip"
127
  },
128
  "qnn_context_binary": {
129
  "tool_versions": {
130
- "qairt": "2.45.0.260326154327"
131
  },
132
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8gen3.zip"
133
  }
134
  },
135
  "qualcomm-snapdragon-x-elite": {
136
  "precompiled_qnn_onnx": {
137
  "tool_versions": {
138
- "qairt": "2.45.0.260326154327",
139
  "onnx_runtime": "1.27.1"
140
  },
141
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_x_elite.zip"
142
  },
143
  "qnn_context_binary": {
144
  "tool_versions": {
145
- "qairt": "2.45.0.260326154327"
146
  },
147
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_x_elite.zip"
148
  }
149
  },
150
  "qualcomm-snapdragon-x2-elite": {
151
  "precompiled_qnn_onnx": {
152
  "tool_versions": {
153
- "qairt": "2.45.0.260326154327",
154
  "onnx_runtime": "1.27.1"
155
  },
156
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_x2_elite.zip"
157
  },
158
  "qnn_context_binary": {
159
  "tool_versions": {
160
- "qairt": "2.45.0.260326154327"
161
  },
162
- "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.62.2/zipformer-qnn_context_binary-float-qualcomm_snapdragon_x2_elite.zip"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
163
  }
164
  }
165
  }
 
1
  {
2
+ "version": "0.63.0",
3
  "precisions": {
4
  "float": {
5
  "chipset_assets": {
6
  "qualcomm-qcs8275": {
7
  "precompiled_qnn_onnx": {
8
  "tool_versions": {
9
+ "qairt": "2.50.0.260828221209",
10
  "onnx_runtime": "1.27.1"
11
  },
12
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs8275.zip"
13
  },
14
  "qnn_context_binary": {
15
  "tool_versions": {
16
+ "qairt": "2.50.0.260828221209"
17
  },
18
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_qcs8275.zip"
19
  }
20
  },
21
  "qualcomm-qcs8550-proxy": {
22
  "precompiled_qnn_onnx": {
23
  "tool_versions": {
24
+ "qairt": "2.50.0.260828221209",
25
  "onnx_runtime": "1.27.1"
26
  },
27
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs8550_proxy.zip"
28
  },
29
  "qnn_context_binary": {
30
  "tool_versions": {
31
+ "qairt": "2.50.0.260828221209"
32
  },
33
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_qcs8550_proxy.zip"
34
  }
35
  },
36
  "qualcomm-qcs9075": {
37
  "precompiled_qnn_onnx": {
38
  "tool_versions": {
39
+ "qairt": "2.50.0.260828221209",
40
  "onnx_runtime": "1.27.1"
41
  },
42
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_qcs9075.zip"
43
  },
44
  "qnn_context_binary": {
45
  "tool_versions": {
46
+ "qairt": "2.50.0.260828221209"
47
  },
48
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_qcs9075.zip"
49
  }
50
  },
51
  "qualcomm-sa7255p": {
52
  "qnn_context_binary": {
53
  "tool_versions": {
54
+ "qairt": "2.50.0.260828221209"
55
  },
56
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_sa7255p.zip"
57
  }
58
  },
59
  "qualcomm-sa8295p": {
60
  "qnn_context_binary": {
61
  "tool_versions": {
62
+ "qairt": "2.50.0.260828221209"
63
  },
64
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_sa8295p.zip"
65
  }
66
  },
67
  "qualcomm-sa8775p": {
68
  "qnn_context_binary": {
69
  "tool_versions": {
70
+ "qairt": "2.50.0.260828221209"
71
  },
72
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_sa8775p.zip"
73
  }
74
  },
75
  "qualcomm-snapdragon-8-elite-for-galaxy": {
76
  "precompiled_qnn_onnx": {
77
  "tool_versions": {
78
+ "qairt": "2.50.0.260828221209",
79
  "onnx_runtime": "1.27.1"
80
  },
81
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_for_galaxy.zip"
82
  },
83
  "qnn_context_binary": {
84
  "tool_versions": {
85
+ "qairt": "2.50.0.260828221209"
86
  },
87
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8_elite_for_galaxy.zip"
88
  }
89
  },
90
  "qualcomm-snapdragon-8-elite-gen5-for-galaxy": {
91
  "precompiled_qnn_onnx": {
92
  "tool_versions": {
93
+ "qairt": "2.50.0.260828221209",
94
  "onnx_runtime": "1.27.1"
95
  },
96
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
97
  },
98
  "qnn_context_binary": {
99
  "tool_versions": {
100
+ "qairt": "2.50.0.260828221209"
101
  },
102
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
103
  }
104
  },
105
  "qualcomm-snapdragon-8gen1": {
106
  "precompiled_qnn_onnx": {
107
  "tool_versions": {
108
+ "qairt": "2.50.0.260828221209",
109
  "onnx_runtime": "1.27.1"
110
  },
111
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen1.zip"
112
  },
113
  "qnn_context_binary": {
114
  "tool_versions": {
115
+ "qairt": "2.50.0.260828221209"
116
  },
117
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8gen1.zip"
118
  }
119
  },
120
  "qualcomm-snapdragon-8gen3": {
121
  "precompiled_qnn_onnx": {
122
  "tool_versions": {
123
+ "qairt": "2.50.0.260828221209",
124
  "onnx_runtime": "1.27.1"
125
  },
126
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_8gen3.zip"
127
  },
128
  "qnn_context_binary": {
129
  "tool_versions": {
130
+ "qairt": "2.50.0.260828221209"
131
  },
132
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_8gen3.zip"
133
  }
134
  },
135
  "qualcomm-snapdragon-x-elite": {
136
  "precompiled_qnn_onnx": {
137
  "tool_versions": {
138
+ "qairt": "2.50.0.260828221209",
139
  "onnx_runtime": "1.27.1"
140
  },
141
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_x_elite.zip"
142
  },
143
  "qnn_context_binary": {
144
  "tool_versions": {
145
+ "qairt": "2.50.0.260828221209"
146
  },
147
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_x_elite.zip"
148
  }
149
  },
150
  "qualcomm-snapdragon-x2-elite": {
151
  "precompiled_qnn_onnx": {
152
  "tool_versions": {
153
+ "qairt": "2.50.0.260828221209",
154
  "onnx_runtime": "1.27.1"
155
  },
156
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-float-qualcomm_snapdragon_x2_elite.zip"
157
  },
158
  "qnn_context_binary": {
159
  "tool_versions": {
160
+ "qairt": "2.50.0.260828221209"
161
  },
162
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-float-qualcomm_snapdragon_x2_elite.zip"
163
+ }
164
+ }
165
+ }
166
+ },
167
+ "mixed": {
168
+ "chipset_assets": {
169
+ "qualcomm-qcs8275": {
170
+ "precompiled_qnn_onnx": {
171
+ "tool_versions": {
172
+ "qairt": "2.50.0.260828221209",
173
+ "onnx_runtime": "1.27.1"
174
+ },
175
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_qcs8275.zip"
176
+ },
177
+ "qnn_context_binary": {
178
+ "tool_versions": {
179
+ "qairt": "2.50.0.260828221209"
180
+ },
181
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_qcs8275.zip"
182
+ }
183
+ },
184
+ "qualcomm-qcs8550-proxy": {
185
+ "precompiled_qnn_onnx": {
186
+ "tool_versions": {
187
+ "qairt": "2.50.0.260828221209",
188
+ "onnx_runtime": "1.27.1"
189
+ },
190
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_qcs8550_proxy.zip"
191
+ },
192
+ "qnn_context_binary": {
193
+ "tool_versions": {
194
+ "qairt": "2.50.0.260828221209"
195
+ },
196
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_qcs8550_proxy.zip"
197
+ }
198
+ },
199
+ "qualcomm-qcs9075": {
200
+ "precompiled_qnn_onnx": {
201
+ "tool_versions": {
202
+ "qairt": "2.50.0.260828221209",
203
+ "onnx_runtime": "1.27.1"
204
+ },
205
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_qcs9075.zip"
206
+ },
207
+ "qnn_context_binary": {
208
+ "tool_versions": {
209
+ "qairt": "2.50.0.260828221209"
210
+ },
211
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_qcs9075.zip"
212
+ }
213
+ },
214
+ "qualcomm-sa7255p": {
215
+ "qnn_context_binary": {
216
+ "tool_versions": {
217
+ "qairt": "2.50.0.260828221209"
218
+ },
219
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_sa7255p.zip"
220
+ }
221
+ },
222
+ "qualcomm-sa8775p": {
223
+ "qnn_context_binary": {
224
+ "tool_versions": {
225
+ "qairt": "2.50.0.260828221209"
226
+ },
227
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_sa8775p.zip"
228
+ }
229
+ },
230
+ "qualcomm-snapdragon-8-elite-for-galaxy": {
231
+ "precompiled_qnn_onnx": {
232
+ "tool_versions": {
233
+ "qairt": "2.50.0.260828221209",
234
+ "onnx_runtime": "1.27.1"
235
+ },
236
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_8_elite_for_galaxy.zip"
237
+ },
238
+ "qnn_context_binary": {
239
+ "tool_versions": {
240
+ "qairt": "2.50.0.260828221209"
241
+ },
242
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_8_elite_for_galaxy.zip"
243
+ }
244
+ },
245
+ "qualcomm-snapdragon-8-elite-gen5-for-galaxy": {
246
+ "precompiled_qnn_onnx": {
247
+ "tool_versions": {
248
+ "qairt": "2.50.0.260828221209",
249
+ "onnx_runtime": "1.27.1"
250
+ },
251
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
252
+ },
253
+ "qnn_context_binary": {
254
+ "tool_versions": {
255
+ "qairt": "2.50.0.260828221209"
256
+ },
257
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_8_elite_gen5_for_galaxy.zip"
258
+ }
259
+ },
260
+ "qualcomm-snapdragon-8gen3": {
261
+ "precompiled_qnn_onnx": {
262
+ "tool_versions": {
263
+ "qairt": "2.50.0.260828221209",
264
+ "onnx_runtime": "1.27.1"
265
+ },
266
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_8gen3.zip"
267
+ },
268
+ "qnn_context_binary": {
269
+ "tool_versions": {
270
+ "qairt": "2.50.0.260828221209"
271
+ },
272
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_8gen3.zip"
273
+ }
274
+ },
275
+ "qualcomm-snapdragon-x-elite": {
276
+ "precompiled_qnn_onnx": {
277
+ "tool_versions": {
278
+ "qairt": "2.50.0.260828221209",
279
+ "onnx_runtime": "1.27.1"
280
+ },
281
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_x_elite.zip"
282
+ },
283
+ "qnn_context_binary": {
284
+ "tool_versions": {
285
+ "qairt": "2.50.0.260828221209"
286
+ },
287
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_x_elite.zip"
288
+ }
289
+ },
290
+ "qualcomm-snapdragon-x2-elite": {
291
+ "precompiled_qnn_onnx": {
292
+ "tool_versions": {
293
+ "qairt": "2.50.0.260828221209",
294
+ "onnx_runtime": "1.27.1"
295
+ },
296
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-precompiled_qnn_onnx-mixed-qualcomm_snapdragon_x2_elite.zip"
297
+ },
298
+ "qnn_context_binary": {
299
+ "tool_versions": {
300
+ "qairt": "2.50.0.260828221209"
301
+ },
302
+ "download_url": "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/zipformer/releases/v0.63.0/zipformer-qnn_context_binary-mixed-qualcomm_snapdragon_x2_elite.zip"
303
  }
304
  }
305
  }