Hugging Face's logo

qualcomm
/
SESR-M5-Quantized

Image-to-Image
PyTorch
LiteRT
ONNX
quantized
android
Model card Files Community
qaihm-bot commited on
Commit
2072746
·
verified ·
1 Parent(s): 4fc8974

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +43 -19
README.md CHANGED
@@ -15,7 +15,7 @@ tags:
15
 
16
  SESR M5 performs efficient on-device upscaling of images.
17
 
18
- This model is an implementation of SESR-M5-Quantized found [here](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/sesr).
19
  This repository provides scripts to run SESR-M5-Quantized on Qualcomm® devices.
20
  More details on model performance across various devices, can be found
21
  [here](https://aihub.qualcomm.com/models/sesr_m5_quantized).
@@ -30,15 +30,35 @@ More details on model performance across various devices, can be found
30
  - Number of parameters: 338K
31
  - Model size: 389 KB
32
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33
 
34
 
35
 
36
- | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
37
- | ---|---|---|---|---|---|---|---|
38
- | Samsung Galaxy S23 Ultra (Android 13) | Snapdragon® 8 Gen 2 | TFLite | 1.332 ms | 0 - 1 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite)
39
- | Samsung Galaxy S23 Ultra (Android 13) | Snapdragon® 8 Gen 2 | QNN Model Library | 0.973 ms | 0 - 4 MB | INT8 | NPU | [SESR-M5-Quantized.so](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.so)
40
-
41
-
42
 
43
  ## Installation
44
 
@@ -94,16 +114,16 @@ device. This script does the following:
94
  ```bash
95
  python -m qai_hub_models.models.sesr_m5_quantized.export
96
  ```
97
-
98
  ```
99
- Profile Job summary of SESR-M5-Quantized
100
- --------------------------------------------------
101
- Device: Snapdragon X Elite CRD (11)
102
- Estimated Inference Time: 0.82 ms
103
- Estimated Peak Memory Range: 0.06-0.06 MB
104
- Compute Units: NPU (25) | Total (25)
105
-
106
-
 
107
  ```
108
 
109
 
@@ -142,15 +162,19 @@ provides instructions on how to use the `.so` shared library in an Android appl
142
  Get more details on SESR-M5-Quantized's performance across various devices [here](https://aihub.qualcomm.com/models/sesr_m5_quantized).
143
  Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
144
 
 
145
  ## License
146
- - The license for the original implementation of SESR-M5-Quantized can be found
147
- [here](https://github.com/quic/aimet-model-zoo/blob/develop/LICENSE.pdf).
148
- - 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)
 
149
 
150
  ## References
151
  * [Collapsible Linear Blocks for Super-Efficient Super Resolution](https://arxiv.org/abs/2103.09404)
152
  * [Source Model Implementation](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/sesr)
153
 
 
 
154
  ## Community
155
  * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
156
  * For questions or feedback please [reach out to us](mailto:[email protected]).
 
15
 
16
  SESR M5 performs efficient on-device upscaling of images.
17
 
18
+ This model is an implementation of SESR-M5-Quantized found [here]({source_repo}).
19
  This repository provides scripts to run SESR-M5-Quantized on Qualcomm® devices.
20
  More details on model performance across various devices, can be found
21
  [here](https://aihub.qualcomm.com/models/sesr_m5_quantized).
 
30
  - Number of parameters: 338K
31
  - Model size: 389 KB
32
 
33
+ | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
34
+ |---|---|---|---|---|---|---|---|---|
35
+ | SESR-M5-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 1.339 ms | 0 - 1 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
36
+ | SESR-M5-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 0.973 ms | 0 - 73 MB | INT8 | NPU | [SESR-M5-Quantized.so](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.so) |
37
+ | SESR-M5-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 1.083 ms | 0 - 2 MB | INT8 | NPU | [SESR-M5-Quantized.onnx](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.onnx) |
38
+ | SESR-M5-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 1.109 ms | 0 - 26 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
39
+ | SESR-M5-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 0.714 ms | 0 - 14 MB | INT8 | NPU | [SESR-M5-Quantized.so](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.so) |
40
+ | SESR-M5-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 0.821 ms | 0 - 30 MB | INT8 | NPU | [SESR-M5-Quantized.onnx](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.onnx) |
41
+ | SESR-M5-Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | TFLITE | 3.597 ms | 0 - 19 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
42
+ | SESR-M5-Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | QNN | 2.908 ms | 0 - 7 MB | INT8 | NPU | Use Export Script |
43
+ | SESR-M5-Quantized | RB5 (Proxy) | QCS8250 Proxy | TFLITE | 19.669 ms | 2 - 4 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
44
+ | SESR-M5-Quantized | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 1.338 ms | 2 - 49 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
45
+ | SESR-M5-Quantized | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 0.684 ms | 0 - 1 MB | INT8 | NPU | Use Export Script |
46
+ | SESR-M5-Quantized | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 1.344 ms | 0 - 8 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
47
+ | SESR-M5-Quantized | SA8255 (Proxy) | SA8255P Proxy | QNN | 0.684 ms | 0 - 1 MB | INT8 | NPU | Use Export Script |
48
+ | SESR-M5-Quantized | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 1.342 ms | 0 - 2 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
49
+ | SESR-M5-Quantized | SA8775 (Proxy) | SA8775P Proxy | QNN | 0.686 ms | 0 - 2 MB | INT8 | NPU | Use Export Script |
50
+ | SESR-M5-Quantized | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 1.351 ms | 0 - 2 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
51
+ | SESR-M5-Quantized | SA8650 (Proxy) | SA8650P Proxy | QNN | 0.687 ms | 0 - 1 MB | INT8 | NPU | Use Export Script |
52
+ | SESR-M5-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 1.985 ms | 2 - 28 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
53
+ | SESR-M5-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 1.106 ms | 0 - 16 MB | INT8 | NPU | Use Export Script |
54
+ | SESR-M5-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 1.351 ms | 0 - 17 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
55
+ | SESR-M5-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 0.599 ms | 0 - 12 MB | INT8 | NPU | Use Export Script |
56
+ | SESR-M5-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 0.592 ms | 0 - 20 MB | INT8 | NPU | [SESR-M5-Quantized.onnx](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.onnx) |
57
+ | SESR-M5-Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 0.798 ms | 0 - 0 MB | INT8 | NPU | Use Export Script |
58
+ | SESR-M5-Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.201 ms | 3 - 3 MB | INT8 | NPU | [SESR-M5-Quantized.onnx](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.onnx) |
59
 
60
 
61
 
 
 
 
 
 
 
62
 
63
  ## Installation
64
 
 
114
  ```bash
115
  python -m qai_hub_models.models.sesr_m5_quantized.export
116
  ```
 
117
  ```
118
+ Profiling Results
119
+ ------------------------------------------------------------
120
+ SESR-M5-Quantized
121
+ Device : Samsung Galaxy S23 (13)
122
+ Runtime : TFLITE
123
+ Estimated inference time (ms) : 1.3
124
+ Estimated peak memory usage (MB): [0, 1]
125
+ Total # Ops : 27
126
+ Compute Unit(s) : NPU (24 ops) CPU (3 ops)
127
  ```
128
 
129
 
 
162
  Get more details on SESR-M5-Quantized's performance across various devices [here](https://aihub.qualcomm.com/models/sesr_m5_quantized).
163
  Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
164
 
165
+
166
  ## License
167
+ * The license for the original implementation of SESR-M5-Quantized can be found [here](https://github.com/quic/aimet-model-zoo/blob/develop/LICENSE.pdf).
168
+ * 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)
169
+
170
+
171
 
172
  ## References
173
  * [Collapsible Linear Blocks for Super-Efficient Super Resolution](https://arxiv.org/abs/2103.09404)
174
  * [Source Model Implementation](https://github.com/quic/aimet-model-zoo/tree/develop/aimet_zoo_torch/sesr)
175
 
176
+
177
+
178
  ## Community
179
  * Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
180
  * For questions or feedback please [reach out to us](mailto:[email protected]).