qaihm-bot commited on
Commit
5842d4a
·
verified ·
1 Parent(s): 9715652

See https://github.com/quic/ai-hub-models/releases/v0.46.1 for changelog.

README.md CHANGED
@@ -10,213 +10,83 @@ pipeline_tag: image-segmentation
10
 
11
  ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/salsanext/web-assets/model_demo.png)
12
 
13
- # SalsaNext: Optimized for Mobile Deployment
14
- ## Semantic segmentation model optimized for LiDAR point cloud data
15
 
16
  SalsaNext is a LiDAR-based model designed for efficient and accurate semantic segmentation.
17
 
18
- This repository provides scripts to run SalsaNext on Qualcomm® devices.
19
- More details on model performance across various devices, can be found
20
- [here](https://aihub.qualcomm.com/models/salsanext).
21
-
22
-
23
-
24
- ### Model Details
25
-
26
- - **Model Type:** Model_use_case.semantic_segmentation
27
- - **Model Stats:**
28
- - Model checkpoint: SalsaNext
29
- - Input resolution: 1x5x64x2048
30
- - Number of parameters: 6.71M
31
- - Model size (float): 25.7 MB
32
-
33
- | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
34
- |---|---|---|---|---|---|---|---|---|
35
- | SalsaNext | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 118.648 ms | 10 - 164 MB | NPU | [SalsaNext.tflite](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.tflite) |
36
- | SalsaNext | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 118.843 ms | 2 - 155 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.dlc) |
37
- | SalsaNext | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 60.783 ms | 10 - 285 MB | NPU | [SalsaNext.tflite](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.tflite) |
38
- | SalsaNext | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 60.893 ms | 3 - 269 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.dlc) |
39
- | SalsaNext | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 33.712 ms | 10 - 12 MB | NPU | [SalsaNext.tflite](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.tflite) |
40
- | SalsaNext | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 34.135 ms | 3 - 5 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.dlc) |
41
- | SalsaNext | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 34.804 ms | 20 - 33 MB | NPU | [SalsaNext.onnx.zip](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.onnx.zip) |
42
- | SalsaNext | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 39.143 ms | 10 - 165 MB | NPU | [SalsaNext.tflite](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.tflite) |
43
- | SalsaNext | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 162.128 ms | 2 - 155 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.dlc) |
44
- | SalsaNext | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 25.538 ms | 8 - 255 MB | NPU | [SalsaNext.tflite](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.tflite) |
45
- | SalsaNext | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 25.214 ms | 3 - 240 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.dlc) |
46
- | SalsaNext | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 26.361 ms | 25 - 240 MB | NPU | [SalsaNext.onnx.zip](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.onnx.zip) |
47
- | SalsaNext | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 21.024 ms | 9 - 164 MB | NPU | [SalsaNext.tflite](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.tflite) |
48
- | SalsaNext | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 20.671 ms | 3 - 161 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.dlc) |
49
- | SalsaNext | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 21.816 ms | 23 - 154 MB | NPU | [SalsaNext.onnx.zip](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.onnx.zip) |
50
- | SalsaNext | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | TFLITE | 15.752 ms | 10 - 169 MB | NPU | [SalsaNext.tflite](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.tflite) |
51
- | SalsaNext | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | QNN_DLC | 15.523 ms | 3 - 161 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.dlc) |
52
- | SalsaNext | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | ONNX | 16.182 ms | 24 - 160 MB | NPU | [SalsaNext.onnx.zip](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.onnx.zip) |
53
- | SalsaNext | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 32.41 ms | 3 - 3 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.dlc) |
54
- | SalsaNext | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 33.248 ms | 33 - 33 MB | NPU | [SalsaNext.onnx.zip](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext.onnx.zip) |
55
- | SalsaNext | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 74.53 ms | 1 - 217 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext_w8a16.dlc) |
56
- | SalsaNext | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 83.038 ms | 1 - 376 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext_w8a16.dlc) |
57
- | SalsaNext | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 40.53 ms | 1 - 4 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext_w8a16.dlc) |
58
- | SalsaNext | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 40.537 ms | 1 - 218 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext_w8a16.dlc) |
59
- | SalsaNext | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 30.535 ms | 1 - 323 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext_w8a16.dlc) |
60
- | SalsaNext | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 24.488 ms | 1 - 229 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext_w8a16.dlc) |
61
- | SalsaNext | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen 5 Mobile | QNN_DLC | 20.052 ms | 1 - 264 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext_w8a16.dlc) |
62
- | SalsaNext | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 40.932 ms | 1 - 1 MB | NPU | [SalsaNext.dlc](https://huggingface.co/qualcomm/SalsaNext/blob/main/SalsaNext_w8a16.dlc) |
63
-
64
-
65
-
66
-
67
- ## Installation
68
-
69
-
70
- Install the package via pip:
71
- ```bash
72
- pip install qai-hub-models
73
- ```
74
-
75
-
76
- ## Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device
77
-
78
- Sign-in to [Qualcomm® AI Hub Workbench](https://workbench.aihub.qualcomm.com/) with your
79
- Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
80
-
81
- With this API token, you can configure your client to run models on the cloud
82
- hosted devices.
83
- ```bash
84
- qai-hub configure --api_token API_TOKEN
85
- ```
86
- Navigate to [docs](https://workbench.aihub.qualcomm.com/docs/) for more information.
87
-
88
-
89
-
90
- ## Demo off target
91
-
92
- The package contains a simple end-to-end demo that downloads pre-trained
93
- weights and runs this model on a sample input.
94
-
95
- ```bash
96
- python -m qai_hub_models.models.salsanext.demo
97
- ```
98
-
99
- The above demo runs a reference implementation of pre-processing, model
100
- inference, and post processing.
101
-
102
- **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
103
- environment, please add the following to your cell (instead of the above).
104
- ```
105
- %run -m qai_hub_models.models.salsanext.demo
106
- ```
107
-
108
-
109
- ### Run model on a cloud-hosted device
110
-
111
- In addition to the demo, you can also run the model on a cloud-hosted Qualcomm®
112
- device. This script does the following:
113
- * Performance check on-device on a cloud-hosted device
114
- * Downloads compiled assets that can be deployed on-device for Android.
115
- * Accuracy check between PyTorch and on-device outputs.
116
-
117
- ```bash
118
- python -m qai_hub_models.models.salsanext.export
119
- ```
120
-
121
-
122
-
123
- ## How does this work?
124
-
125
- This [export script](https://aihub.qualcomm.com/models/salsanext/qai_hub_models/models/SalsaNext/export.py)
126
- leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) to optimize, validate, and deploy this model
127
- on-device. Lets go through each step below in detail:
128
-
129
- Step 1: **Compile model for on-device deployment**
130
-
131
- To compile a PyTorch model for on-device deployment, we first trace the model
132
- in memory using the `jit.trace` and then call the `submit_compile_job` API.
133
-
134
- ```python
135
- import torch
136
-
137
- import qai_hub as hub
138
- from qai_hub_models.models.salsanext import Model
139
-
140
- # Load the model
141
- torch_model = Model.from_pretrained()
142
-
143
- # Device
144
- device = hub.Device("Samsung Galaxy S25")
145
-
146
- # Trace model
147
- input_shape = torch_model.get_input_spec()
148
- sample_inputs = torch_model.sample_inputs()
149
-
150
- pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
151
-
152
- # Compile model on a specific device
153
- compile_job = hub.submit_compile_job(
154
- model=pt_model,
155
- device=device,
156
- input_specs=torch_model.get_input_spec(),
157
- )
158
-
159
- # Get target model to run on-device
160
- target_model = compile_job.get_target_model()
161
-
162
- ```
163
-
164
-
165
- Step 2: **Performance profiling on cloud-hosted device**
166
-
167
- After compiling models from step 1. Models can be profiled model on-device using the
168
- `target_model`. Note that this scripts runs the model on a device automatically
169
- provisioned in the cloud. Once the job is submitted, you can navigate to a
170
- provided job URL to view a variety of on-device performance metrics.
171
- ```python
172
- profile_job = hub.submit_profile_job(
173
- model=target_model,
174
- device=device,
175
- )
176
-
177
- ```
178
-
179
- Step 3: **Verify on-device accuracy**
180
-
181
- To verify the accuracy of the model on-device, you can run on-device inference
182
- on sample input data on the same cloud hosted device.
183
- ```python
184
- input_data = torch_model.sample_inputs()
185
- inference_job = hub.submit_inference_job(
186
- model=target_model,
187
- device=device,
188
- inputs=input_data,
189
- )
190
- on_device_output = inference_job.download_output_data()
191
-
192
- ```
193
- With the output of the model, you can compute like PSNR, relative errors or
194
- spot check the output with expected output.
195
-
196
- **Note**: This on-device profiling and inference requires access to Qualcomm®
197
- AI Hub Workbench. [Sign up for access](https://myaccount.qualcomm.com/signup).
198
-
199
-
200
-
201
-
202
- ## Deploying compiled model to Android
203
-
204
-
205
- The models can be deployed using multiple runtimes:
206
- - TensorFlow Lite (`.tflite` export): [This
207
- tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a
208
- guide to deploy the .tflite model in an Android application.
209
-
210
-
211
- - QNN (`.so` export ): This [sample
212
- app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html)
213
- provides instructions on how to use the `.so` shared library in an Android application.
214
-
215
-
216
- ## View on Qualcomm® AI Hub
217
- Get more details on SalsaNext's performance across various devices [here](https://aihub.qualcomm.com/models/salsanext).
218
- Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
219
-
220
 
221
  ## License
222
  * The license for the original implementation of SalsaNext can be found
@@ -224,9 +94,6 @@ Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
224
 
225
 
226
 
227
-
228
  ## Community
229
  * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
230
  * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).
231
-
232
-
 
10
 
11
  ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/salsanext/web-assets/model_demo.png)
12
 
13
+ # SalsaNext: Optimized for Qualcomm Devices
 
14
 
15
  SalsaNext is a LiDAR-based model designed for efficient and accurate semantic segmentation.
16
 
17
+ This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/salsanext) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary).
18
+
19
+ 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.
20
+
21
+ ## Getting Started
22
+ There are two ways to deploy this model on your device:
23
+
24
+ ### Option 1: Download Pre-Exported Models
25
+
26
+ Below are pre-exported model assets ready for deployment.
27
+
28
+ | Runtime | Precision | Chipset | SDK Versions | Download |
29
+ |---|---|---|---|---|
30
+ | ONNX | float | Universal | QAIRT 2.37, ONNX Runtime 1.23.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/salsanext/releases/v0.46.1/salsanext-onnx-float.zip)
31
+ | QNN_DLC | float | Universal | QAIRT 2.42 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/salsanext/releases/v0.46.1/salsanext-qnn_dlc-float.zip)
32
+ | QNN_DLC | w8a16 | Universal | QAIRT 2.42 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/salsanext/releases/v0.46.1/salsanext-qnn_dlc-w8a16.zip)
33
+ | TFLITE | float | Universal | QAIRT 2.42, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/salsanext/releases/v0.46.1/salsanext-tflite-float.zip)
34
+
35
+ For more device-specific assets and performance metrics, visit **[SalsaNext on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/salsanext)**.
36
+
37
+
38
+ ### Option 2: Export with Custom Configurations
39
+
40
+ Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/salsanext) Python library to compile and export the model with your own:
41
+ - Custom weights (e.g., fine-tuned checkpoints)
42
+ - Custom input shapes
43
+ - Target device and runtime configurations
44
+
45
+ This option is ideal if you need to customize the model beyond the default configuration provided here.
46
+
47
+ See our repository for [SalsaNext on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/salsanext) for usage instructions.
48
+
49
+ ## Model Details
50
+
51
+ **Model Type:** Model_use_case.semantic_segmentation
52
+
53
+ **Model Stats:**
54
+ - Model checkpoint: SalsaNext
55
+ - Input resolution: 1x5x64x2048
56
+ - Number of parameters: 6.71M
57
+ - Model size (float): 25.7 MB
58
+
59
+ ## Performance Summary
60
+ | Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
61
+ |---|---|---|---|---|---|---
62
+ | SalsaNext | ONNX | float | Snapdragon® X Elite | 33.24 ms | 33 - 33 MB | NPU
63
+ | SalsaNext | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 27.18 ms | 27 - 241 MB | NPU
64
+ | SalsaNext | ONNX | float | Qualcomm® QCS8550 (Proxy) | 34.597 ms | 20 - 40 MB | NPU
65
+ | SalsaNext | ONNX | float | Qualcomm® QCS9075 | 40.745 ms | 23 - 25 MB | NPU
66
+ | SalsaNext | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 21.369 ms | 23 - 155 MB | NPU
67
+ | SalsaNext | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 15.944 ms | 26 - 164 MB | NPU
68
+ | SalsaNext | QNN_DLC | float | Snapdragon® X Elite | 30.095 ms | 3 - 3 MB | NPU
69
+ | SalsaNext | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 22.357 ms | 0 - 261 MB | NPU
70
+ | SalsaNext | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 116.123 ms | 1 - 176 MB | NPU
71
+ | SalsaNext | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 30.828 ms | 3 - 5 MB | NPU
72
+ | SalsaNext | QNN_DLC | float | Qualcomm® QCS9075 | 40.105 ms | 3 - 17 MB | NPU
73
+ | SalsaNext | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 55.424 ms | 0 - 296 MB | NPU
74
+ | SalsaNext | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 20.556 ms | 3 - 186 MB | NPU
75
+ | SalsaNext | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 15.21 ms | 3 - 191 MB | NPU
76
+ | SalsaNext | QNN_DLC | w8a16 | Snapdragon® X Elite | 43.137 ms | 1 - 1 MB | NPU
77
+ | SalsaNext | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 34.328 ms | 0 - 365 MB | NPU
78
+ | SalsaNext | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 44.739 ms | 1 - 4 MB | NPU
79
+ | SalsaNext | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 50.506 ms | 1 - 9 MB | NPU
80
+ | SalsaNext | QNN_DLC | w8a16 | Qualcomm® QCS8450 (Proxy) | 82.609 ms | 1 - 424 MB | NPU
81
+ | SalsaNext | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 28.297 ms | 1 - 275 MB | NPU
82
+ | SalsaNext | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 19.552 ms | 1 - 314 MB | NPU
83
+ | SalsaNext | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 22.149 ms | 8 - 277 MB | NPU
84
+ | SalsaNext | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 613.33 ms | 8 - 37 MB | GPU
85
+ | SalsaNext | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 30.525 ms | 0 - 17 MB | NPU
86
+ | SalsaNext | TFLITE | float | Qualcomm® QCS9075 | 39.65 ms | 8 - 36 MB | NPU
87
+ | SalsaNext | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 56.073 ms | 10 - 310 MB | NPU
88
+ | SalsaNext | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 18.097 ms | 9 - 192 MB | NPU
89
+ | SalsaNext | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 15.303 ms | 10 - 198 MB | NPU
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
90
 
91
  ## License
92
  * The license for the original implementation of SalsaNext can be found
 
94
 
95
 
96
 
 
97
  ## Community
98
  * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
99
  * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).
 
 
SalsaNext_float.dlc DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:deda3665ea2aae10c73ed402235935ced4bbf6484732102beb2e477ae5bbf1e3
3
- size 27059644
 
 
 
 
SalsaNext_float.onnx.zip DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:be13d2a9bb25a48529c48d6c0806c2592021f59561b51f20cd18076b9e84c421
3
- size 25068906
 
 
 
 
SalsaNext_float.tflite DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:caf8b11b9e069ce5afdfe3652d0d1541eb2a8aa5ecc849600e0255ea040097a6
3
- size 26904848
 
 
 
 
SalsaNext_w8a16.dlc DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:3cb08679ccf5f079915b5b919c938f33a901dcfc1e62f974e94108eb6e0cfa76
3
- size 7391140
 
 
 
 
tool-versions.yaml DELETED
@@ -1,3 +0,0 @@
1
- tool_versions:
2
- qnn_dlc:
3
- qairt: 2.41.0.251128145156_191518