Hugging Face's logo

qualcomm
/
Segment-Anything-Model-2

Image Segmentation
PyTorch
LiteRT
foundation
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
3b28115
·
verified ·
1 Parent(s): 3c59dc0

v0.31.0

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.31.0 for changelog.

Files changed (6) hide show
  1. .gitattributes +2 -0
  2. README.md +286 -0
  3. Segment-Anything-Model-2_SAM2Decoder.dlc +3 -0
  4. Segment-Anything-Model-2_SAM2Decoder.tflite +3 -0
  5. Segment-Anything-Model-2_SAM2Encoder.dlc +3 -0
  6. Segment-Anything-Model-2_SAM2Encoder.tflite +3 -0
.gitattributes CHANGED
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ Segment-Anything-Model-2_SAM2Decoder.dlc filter=lfs diff=lfs merge=lfs -text
37
+ Segment-Anything-Model-2_SAM2Encoder.dlc filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,286 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: pytorch
3
+ license: other
4
+ tags:
5
+ - foundation
6
+ - android
7
+ pipeline_tag: image-segmentation
8
+
9
+ ---
10
+
11
+ ![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/sam2/web-assets/model_demo.png)
12
+
13
+ # Segment-Anything-Model-2: Optimized for Mobile Deployment
14
+ ## High-quality segmentation in images and videos with real-time performance and minimal user interaction
15
+
16
+
17
+ SAM 2, the successor to Meta's Segment Anything Model (SAM), is a cutting-edge tool designed for comprehensive object segmentation in both images and videos. It excels in handling complex visual data through a unified, promptable model architecture that supports real-time processing and zero-shot generalization.
18
+
19
+ This model is an implementation of Segment-Anything-Model-2 found [here](https://github.com/facebookresearch/sam2).
20
+
21
+
22
+ This repository provides scripts to run Segment-Anything-Model-2 on Qualcomm® devices.
23
+ More details on model performance across various devices, can be found
24
+ [here](https://aihub.qualcomm.com/models/sam2).
25
+
26
+
27
+ ### Model Details
28
+
29
+ - **Model Type:** Model_use_case.semantic_segmentation
30
+ - **Model Stats:**
31
+ - Model checkpoint: sam2.1_hiera_t
32
+ - Input resolution: 720p (720x1280)
33
+ - Number of parameters (SAM2Encoder): 33.5M
34
+ - Model size (SAM2Encoder): 128 MB
35
+ - Number of parameters (SAM2Decoder): 6.22M
36
+ - Model size (SAM2Decoder): 23.7 MB
37
+
38
+ | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
39
+ |---|---|---|---|---|---|---|---|---|
40
+ | SAM2Encoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 609.767 ms | 92 - 296 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
41
+ | SAM2Encoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 367.288 ms | 85 - 296 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
42
+ | SAM2Encoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 384.843 ms | 12 - 2403 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
43
+ | SAM2Encoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 252.662 ms | 92 - 117 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
44
+ | SAM2Encoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 212.081 ms | 15 - 86 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
45
+ | SAM2Encoder | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 290.379 ms | 92 - 295 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
46
+ | SAM2Encoder | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 609.767 ms | 92 - 296 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
47
+ | SAM2Encoder | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 259.454 ms | 92 - 117 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
48
+ | SAM2Encoder | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 214.113 ms | 12 - 93 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
49
+ | SAM2Encoder | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 380.286 ms | 92 - 307 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
50
+ | SAM2Encoder | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 263.816 ms | 92 - 117 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
51
+ | SAM2Encoder | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 214.315 ms | 12 - 89 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
52
+ | SAM2Encoder | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 290.379 ms | 92 - 295 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
53
+ | SAM2Encoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 265.703 ms | 92 - 115 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
54
+ | SAM2Encoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 215.128 ms | 12 - 87 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
55
+ | SAM2Encoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 191.967 ms | 92 - 308 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
56
+ | SAM2Encoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 178.194 ms | 90 - 298 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
57
+ | SAM2Encoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 139.549 ms | 12 - 1333 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
58
+ | SAM2Encoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 260.348 ms | 800 - 800 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
59
+ | SAM2Decoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 17.97 ms | 0 - 54 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
60
+ | SAM2Decoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 10.167 ms | 0 - 57 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
61
+ | SAM2Decoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 11.324 ms | 15 - 76 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
62
+ | SAM2Decoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 8.133 ms | 0 - 30 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
63
+ | SAM2Decoder | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 6.962 ms | 16 - 35 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
64
+ | SAM2Decoder | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 9.556 ms | 0 - 55 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
65
+ | SAM2Decoder | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 17.97 ms | 0 - 54 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
66
+ | SAM2Decoder | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 8.173 ms | 0 - 33 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
67
+ | SAM2Decoder | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 6.96 ms | 16 - 39 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
68
+ | SAM2Decoder | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 11.752 ms | 0 - 52 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
69
+ | SAM2Decoder | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 8.166 ms | 0 - 32 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
70
+ | SAM2Decoder | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 6.987 ms | 16 - 37 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
71
+ | SAM2Decoder | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 9.556 ms | 0 - 55 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
72
+ | SAM2Decoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 8.209 ms | 0 - 33 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
73
+ | SAM2Decoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 6.982 ms | 15 - 38 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
74
+ | SAM2Decoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 5.59 ms | 0 - 65 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
75
+ | SAM2Decoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 5.651 ms | 0 - 56 MB | NPU | [Segment-Anything-Model-2.tflite](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.tflite) |
76
+ | SAM2Decoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 4.238 ms | 8 - 62 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
77
+ | SAM2Decoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 7.652 ms | 16 - 16 MB | NPU | [Segment-Anything-Model-2.dlc](https://huggingface.co/qualcomm/Segment-Anything-Model-2/blob/main/Segment-Anything-Model-2.dlc) |
78
+
79
+
80
+
81
+
82
+ ## Installation
83
+
84
+
85
+ Install the package via pip:
86
+ ```bash
87
+ pip install "qai-hub-models[sam2]"
88
+ ```
89
+
90
+
91
+ ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
92
+
93
+ Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
94
+ Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
95
+
96
+ With this API token, you can configure your client to run models on the cloud
97
+ hosted devices.
98
+ ```bash
99
+ qai-hub configure --api_token API_TOKEN
100
+ ```
101
+ Navigate to [docs](https://app.aihub.qualcomm.com/docs/) for more information.
102
+
103
+
104
+
105
+ ## Demo off target
106
+
107
+ The package contains a simple end-to-end demo that downloads pre-trained
108
+ weights and runs this model on a sample input.
109
+
110
+ ```bash
111
+ python -m qai_hub_models.models.sam2.demo
112
+ ```
113
+
114
+ The above demo runs a reference implementation of pre-processing, model
115
+ inference, and post processing.
116
+
117
+ **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
118
+ environment, please add the following to your cell (instead of the above).
119
+ ```
120
+ %run -m qai_hub_models.models.sam2.demo
121
+ ```
122
+
123
+
124
+ ### Run model on a cloud-hosted device
125
+
126
+ In addition to the demo, you can also run the model on a cloud-hosted Qualcomm®
127
+ device. This script does the following:
128
+ * Performance check on-device on a cloud-hosted device
129
+ * Downloads compiled assets that can be deployed on-device for Android.
130
+ * Accuracy check between PyTorch and on-device outputs.
131
+
132
+ ```bash
133
+ python -m qai_hub_models.models.sam2.export
134
+ ```
135
+ ```
136
+ Profiling Results
137
+ ------------------------------------------------------------
138
+ SAM2Encoder
139
+ Device : cs_8275 (ANDROID 14)
140
+ Runtime : TFLITE
141
+ Estimated inference time (ms) : 609.8
142
+ Estimated peak memory usage (MB): [92, 296]
143
+ Total # Ops : 647
144
+ Compute Unit(s) : npu (593 ops) gpu (0 ops) cpu (54 ops)
145
+
146
+ ------------------------------------------------------------
147
+ SAM2Decoder
148
+ Device : cs_8275 (ANDROID 14)
149
+ Runtime : TFLITE
150
+ Estimated inference time (ms) : 18.0
151
+ Estimated peak memory usage (MB): [0, 54]
152
+ Total # Ops : 889
153
+ Compute Unit(s) : npu (889 ops) gpu (0 ops) cpu (0 ops)
154
+ ```
155
+
156
+
157
+ ## How does this work?
158
+
159
+ This [export script](https://aihub.qualcomm.com/models/sam2/qai_hub_models/models/Segment-Anything-Model-2/export.py)
160
+ leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) to optimize, validate, and deploy this model
161
+ on-device. Lets go through each step below in detail:
162
+
163
+ Step 1: **Compile model for on-device deployment**
164
+
165
+ To compile a PyTorch model for on-device deployment, we first trace the model
166
+ in memory using the `jit.trace` and then call the `submit_compile_job` API.
167
+
168
+ ```python
169
+ import torch
170
+
171
+ import qai_hub as hub
172
+ from qai_hub_models.models.sam2 import Model
173
+
174
+ # Load the model
175
+ torch_model = Model.from_pretrained()
176
+
177
+ # Device
178
+ device = hub.Device("Samsung Galaxy S24")
179
+
180
+ # Trace model
181
+ input_shape = torch_model.get_input_spec()
182
+ sample_inputs = torch_model.sample_inputs()
183
+
184
+ pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
185
+
186
+ # Compile model on a specific device
187
+ compile_job = hub.submit_compile_job(
188
+ model=pt_model,
189
+ device=device,
190
+ input_specs=torch_model.get_input_spec(),
191
+ )
192
+
193
+ # Get target model to run on-device
194
+ target_model = compile_job.get_target_model()
195
+
196
+ ```
197
+
198
+
199
+ Step 2: **Performance profiling on cloud-hosted device**
200
+
201
+ After compiling models from step 1. Models can be profiled model on-device using the
202
+ `target_model`. Note that this scripts runs the model on a device automatically
203
+ provisioned in the cloud. Once the job is submitted, you can navigate to a
204
+ provided job URL to view a variety of on-device performance metrics.
205
+ ```python
206
+ profile_job = hub.submit_profile_job(
207
+ model=target_model,
208
+ device=device,
209
+ )
210
+
211
+ ```
212
+
213
+ Step 3: **Verify on-device accuracy**
214
+
215
+ To verify the accuracy of the model on-device, you can run on-device inference
216
+ on sample input data on the same cloud hosted device.
217
+ ```python
218
+ input_data = torch_model.sample_inputs()
219
+ inference_job = hub.submit_inference_job(
220
+ model=target_model,
221
+ device=device,
222
+ inputs=input_data,
223
+ )
224
+ on_device_output = inference_job.download_output_data()
225
+
226
+ ```
227
+ With the output of the model, you can compute like PSNR, relative errors or
228
+ spot check the output with expected output.
229
+
230
+ **Note**: This on-device profiling and inference requires access to Qualcomm®
231
+ AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
232
+
233
+
234
+
235
+ ## Run demo on a cloud-hosted device
236
+
237
+ You can also run the demo on-device.
238
+
239
+ ```bash
240
+ python -m qai_hub_models.models.sam2.demo --eval-mode on-device
241
+ ```
242
+
243
+ **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
244
+ environment, please add the following to your cell (instead of the above).
245
+ ```
246
+ %run -m qai_hub_models.models.sam2.demo -- --eval-mode on-device
247
+ ```
248
+
249
+
250
+ ## Deploying compiled model to Android
251
+
252
+
253
+ The models can be deployed using multiple runtimes:
254
+ - TensorFlow Lite (`.tflite` export): [This
255
+ tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a
256
+ guide to deploy the .tflite model in an Android application.
257
+
258
+
259
+ - QNN (`.so` export ): This [sample
260
+ app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html)
261
+ provides instructions on how to use the `.so` shared library in an Android application.
262
+
263
+
264
+ ## View on Qualcomm® AI Hub
265
+ Get more details on Segment-Anything-Model-2's performance across various devices [here](https://aihub.qualcomm.com/models/sam2).
266
+ Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
267
+
268
+
269
+ ## License
270
+ * The license for the original implementation of Segment-Anything-Model-2 can be found
271
+ [here](https://github.com/facebookresearch/sam2/blob/main/LICENSE).
272
+ * The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf)
273
+
274
+
275
+
276
+ ## References
277
+ * [SAM 2 Segment Anything in Images and Videos](https://arxiv.org/abs/2408.00714)
278
+ * [Source Model Implementation](https://github.com/facebookresearch/sam2)
279
+
280
+
281
+
282
+ ## Community
283
+ * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
284
+ * For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).
285
+
286
+
Segment-Anything-Model-2_SAM2Decoder.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6627a24f888aa8215154652eec310890dba83361be619ccfc973a04d6680be8f
3
+ size 25337779
Segment-Anything-Model-2_SAM2Decoder.tflite ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:36b8af9046e6cfd5f3a004e3810f57caf8b14899d67b77ca12d81762f07c1729
3
+ size 24878352
Segment-Anything-Model-2_SAM2Encoder.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:781f5e13b85c2372a386f0daef8aadf0544c8cde19acf87f98456382c5b21756
3
+ size 134437963
Segment-Anything-Model-2_SAM2Encoder.tflite ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:328be9ab2563d70bc417240b5e5c5d86b716047fbccc9662184529a3eb0487bd
3
+ size 134225260