Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
@@ -35,38 +35,34 @@ More details on model performance across various devices, can be found
|
|
35 |
|
36 |
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
|
37 |
|---|---|---|---|---|---|---|---|---|
|
38 |
-
| SESR-M5
|
39 |
-
| SESR-M5
|
40 |
-
| SESR-M5
|
41 |
-
| SESR-M5
|
42 |
-
| SESR-M5
|
43 |
-
| SESR-M5
|
44 |
-
| SESR-M5
|
45 |
-
| SESR-M5
|
46 |
-
| SESR-M5
|
47 |
-
| SESR-M5
|
48 |
-
| SESR-M5
|
49 |
-
| SESR-M5
|
50 |
-
| SESR-M5
|
51 |
-
| SESR-M5
|
52 |
-
| SESR-M5
|
53 |
-
| SESR-M5
|
54 |
-
| SESR-M5
|
55 |
-
| SESR-M5
|
56 |
-
| SESR-M5
|
57 |
-
| SESR-M5
|
58 |
-
| SESR-M5
|
59 |
-
| SESR-M5
|
60 |
-
| SESR-M5
|
61 |
-
| SESR-M5
|
62 |
-
| SESR-M5
|
63 |
-
| SESR-M5
|
64 |
-
| SESR-M5
|
65 |
-
| SESR-M5
|
66 |
-
| SESR-M5-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 2.121 ms | 2 - 28 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.tflite) |
|
67 |
-
| SESR-M5-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 1.003 ms | 0 - 26 MB | INT8 | NPU | Use Export Script |
|
68 |
-
| SESR-M5-Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 0.782 ms | 0 - 0 MB | INT8 | NPU | Use Export Script |
|
69 |
-
| SESR-M5-Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.048 ms | 3 - 3 MB | INT8 | NPU | [SESR-M5-Quantized.onnx](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5-Quantized.onnx) |
|
70 |
|
71 |
|
72 |
|
@@ -127,10 +123,10 @@ python -m qai_hub_models.models.sesr_m5_quantized.export
|
|
127 |
```
|
128 |
Profiling Results
|
129 |
------------------------------------------------------------
|
130 |
-
SESR-M5
|
131 |
Device : Samsung Galaxy S23 (13)
|
132 |
Runtime : TFLITE
|
133 |
-
Estimated inference time (ms) : 1.
|
134 |
Estimated peak memory usage (MB): [0, 6]
|
135 |
Total # Ops : 27
|
136 |
Compute Unit(s) : NPU (24 ops) CPU (3 ops)
|
@@ -215,20 +211,6 @@ AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
|
|
215 |
|
216 |
|
217 |
|
218 |
-
## Run demo on a cloud-hosted device
|
219 |
-
|
220 |
-
You can also run the demo on-device.
|
221 |
-
|
222 |
-
```bash
|
223 |
-
python -m qai_hub_models.models.sesr_m5_quantized.demo --on-device
|
224 |
-
```
|
225 |
-
|
226 |
-
**NOTE**: If you want running in a Jupyter Notebook or Google Colab like
|
227 |
-
environment, please add the following to your cell (instead of the above).
|
228 |
-
```
|
229 |
-
%run -m qai_hub_models.models.sesr_m5_quantized.demo -- --on-device
|
230 |
-
```
|
231 |
-
|
232 |
|
233 |
## Deploying compiled model to Android
|
234 |
|
|
|
35 |
|
36 |
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
|
37 |
|---|---|---|---|---|---|---|---|---|
|
38 |
+
| SESR-M5 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 1.387 ms | 0 - 6 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
39 |
+
| SESR-M5 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 0.642 ms | 0 - 2 MB | INT8 | NPU | [SESR-M5-Quantized.so](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.so) |
|
40 |
+
| SESR-M5 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 0.982 ms | 0 - 8 MB | INT8 | NPU | [SESR-M5-Quantized.onnx](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.onnx) |
|
41 |
+
| SESR-M5 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 1.104 ms | 0 - 29 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
42 |
+
| SESR-M5 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 0.448 ms | 0 - 20 MB | INT8 | NPU | [SESR-M5-Quantized.so](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.so) |
|
43 |
+
| SESR-M5 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 0.656 ms | 0 - 23 MB | INT8 | NPU | [SESR-M5-Quantized.onnx](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.onnx) |
|
44 |
+
| SESR-M5 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 1.578 ms | 0 - 20 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
45 |
+
| SESR-M5 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 0.611 ms | 0 - 20 MB | INT8 | NPU | Use Export Script |
|
46 |
+
| SESR-M5 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 0.703 ms | 0 - 20 MB | INT8 | NPU | [SESR-M5-Quantized.onnx](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.onnx) |
|
47 |
+
| SESR-M5 | SA7255P ADP | SA7255P | TFLITE | 14.881 ms | 0 - 12 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
48 |
+
| SESR-M5 | SA7255P ADP | SA7255P | QNN | 10.531 ms | 0 - 9 MB | INT8 | NPU | Use Export Script |
|
49 |
+
| SESR-M5 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 1.346 ms | 0 - 6 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
50 |
+
| SESR-M5 | SA8255 (Proxy) | SA8255P Proxy | QNN | 0.624 ms | 0 - 2 MB | INT8 | NPU | Use Export Script |
|
51 |
+
| SESR-M5 | SA8295P ADP | SA8295P | TFLITE | 2.886 ms | 2 - 20 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
52 |
+
| SESR-M5 | SA8295P ADP | SA8295P | QNN | 1.364 ms | 0 - 17 MB | INT8 | NPU | Use Export Script |
|
53 |
+
| SESR-M5 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 1.352 ms | 0 - 7 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
54 |
+
| SESR-M5 | SA8650 (Proxy) | SA8650P Proxy | QNN | 0.627 ms | 0 - 2 MB | INT8 | NPU | Use Export Script |
|
55 |
+
| SESR-M5 | RB3 Gen 2 (Proxy) | QCS6490 Proxy | TFLITE | 4.728 ms | 1 - 21 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
56 |
+
| SESR-M5 | RB3 Gen 2 (Proxy) | QCS6490 Proxy | QNN | 2.946 ms | 0 - 11 MB | INT8 | NPU | Use Export Script |
|
57 |
+
| SESR-M5 | RB5 (Proxy) | QCS8250 Proxy | TFLITE | 21.371 ms | 2 - 4 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
58 |
+
| SESR-M5 | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 14.881 ms | 0 - 12 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
59 |
+
| SESR-M5 | QCS8275 (Proxy) | QCS8275 Proxy | QNN | 10.531 ms | 0 - 9 MB | INT8 | NPU | Use Export Script |
|
60 |
+
| SESR-M5 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 1.352 ms | 0 - 6 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
61 |
+
| SESR-M5 | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 0.626 ms | 0 - 10 MB | INT8 | NPU | Use Export Script |
|
62 |
+
| SESR-M5 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 2.761 ms | 0 - 24 MB | INT8 | NPU | [SESR-M5-Quantized.tflite](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.tflite) |
|
63 |
+
| SESR-M5 | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 0.99 ms | 0 - 22 MB | INT8 | NPU | Use Export Script |
|
64 |
+
| SESR-M5 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 0.763 ms | 0 - 0 MB | INT8 | NPU | Use Export Script |
|
65 |
+
| SESR-M5 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 0.966 ms | 3 - 3 MB | INT8 | NPU | [SESR-M5-Quantized.onnx](https://huggingface.co/qualcomm/SESR-M5-Quantized/blob/main/SESR-M5.onnx) |
|
|
|
|
|
|
|
|
|
66 |
|
67 |
|
68 |
|
|
|
123 |
```
|
124 |
Profiling Results
|
125 |
------------------------------------------------------------
|
126 |
+
SESR-M5
|
127 |
Device : Samsung Galaxy S23 (13)
|
128 |
Runtime : TFLITE
|
129 |
+
Estimated inference time (ms) : 1.4
|
130 |
Estimated peak memory usage (MB): [0, 6]
|
131 |
Total # Ops : 27
|
132 |
Compute Unit(s) : NPU (24 ops) CPU (3 ops)
|
|
|
211 |
|
212 |
|
213 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
214 |
|
215 |
## Deploying compiled model to Android
|
216 |
|