v0.30.5
Browse filesSee https://github.com/quic/ai-hub-models/releases/v0.30.5 for changelog.
README.md
CHANGED
|
@@ -34,35 +34,35 @@ More details on model performance across various devices, can be found
|
|
| 34 |
|
| 35 |
| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
|
| 36 |
|---|---|---|---|---|---|---|---|---|
|
| 37 |
-
| Depth-Anything-V2 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE |
|
| 38 |
-
| Depth-Anything-V2 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN |
|
| 39 |
-
| Depth-Anything-V2 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE |
|
| 40 |
-
| Depth-Anything-V2 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN |
|
| 41 |
-
| Depth-Anything-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE |
|
| 42 |
-
| Depth-Anything-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN |
|
| 43 |
-
| Depth-Anything-V2 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE |
|
| 44 |
-
| Depth-Anything-V2 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN |
|
| 45 |
-
| Depth-Anything-V2 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE |
|
| 46 |
-
| Depth-Anything-V2 | float | SA7255P ADP | Qualcomm® SA7255P | QNN |
|
| 47 |
-
| Depth-Anything-V2 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE |
|
| 48 |
-
| Depth-Anything-V2 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN |
|
| 49 |
-
| Depth-Anything-V2 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE |
|
| 50 |
-
| Depth-Anything-V2 | float | SA8295P ADP | Qualcomm® SA8295P | QNN |
|
| 51 |
-
| Depth-Anything-V2 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE |
|
| 52 |
-
| Depth-Anything-V2 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN |
|
| 53 |
-
| Depth-Anything-V2 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE |
|
| 54 |
-
| Depth-Anything-V2 | float | SA8775P ADP | Qualcomm® SA8775P | QNN |
|
| 55 |
-
| Depth-Anything-V2 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE |
|
| 56 |
-
| Depth-Anything-V2 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN |
|
| 57 |
-
| Depth-Anything-V2 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX |
|
| 58 |
-
| Depth-Anything-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE |
|
| 59 |
-
| Depth-Anything-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN |
|
| 60 |
-
| Depth-Anything-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX |
|
| 61 |
-
| Depth-Anything-V2 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE |
|
| 62 |
-
| Depth-Anything-V2 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN |
|
| 63 |
-
| Depth-Anything-V2 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX |
|
| 64 |
-
| Depth-Anything-V2 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN |
|
| 65 |
-
| Depth-Anything-V2 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX |
|
| 66 |
|
| 67 |
|
| 68 |
|
|
@@ -126,8 +126,8 @@ Profiling Results
|
|
| 126 |
Depth-Anything-V2
|
| 127 |
Device : cs_8275 (ANDROID 14)
|
| 128 |
Runtime : TFLITE
|
| 129 |
-
Estimated inference time (ms) :
|
| 130 |
-
Estimated peak memory usage (MB): [1,
|
| 131 |
Total # Ops : 646
|
| 132 |
Compute Unit(s) : npu (646 ops) gpu (0 ops) cpu (0 ops)
|
| 133 |
```
|
|
@@ -216,13 +216,13 @@ AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
|
|
| 216 |
You can also run the demo on-device.
|
| 217 |
|
| 218 |
```bash
|
| 219 |
-
python -m qai_hub_models.models.depth_anything_v2.demo --on-device
|
| 220 |
```
|
| 221 |
|
| 222 |
**NOTE**: If you want running in a Jupyter Notebook or Google Colab like
|
| 223 |
environment, please add the following to your cell (instead of the above).
|
| 224 |
```
|
| 225 |
-
%run -m qai_hub_models.models.depth_anything_v2.demo -- --on-device
|
| 226 |
```
|
| 227 |
|
| 228 |
|
|
|
|
| 34 |
|
| 35 |
| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
|
| 36 |
|---|---|---|---|---|---|---|---|---|
|
| 37 |
+
| Depth-Anything-V2 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 889.518 ms | 1 - 598 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 38 |
+
| Depth-Anything-V2 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN | 245.901 ms | 2 - 10 MB | NPU | Use Export Script |
|
| 39 |
+
| Depth-Anything-V2 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 192.834 ms | 1 - 628 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 40 |
+
| Depth-Anything-V2 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN | 244.88 ms | 1 - 669 MB | NPU | Use Export Script |
|
| 41 |
+
| Depth-Anything-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 144.789 ms | 1 - 104 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 42 |
+
| Depth-Anything-V2 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN | 131.932 ms | 3 - 5 MB | NPU | Use Export Script |
|
| 43 |
+
| Depth-Anything-V2 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 143.826 ms | 1 - 597 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 44 |
+
| Depth-Anything-V2 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN | 130.744 ms | 1 - 11 MB | NPU | Use Export Script |
|
| 45 |
+
| Depth-Anything-V2 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 889.518 ms | 1 - 598 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 46 |
+
| Depth-Anything-V2 | float | SA7255P ADP | Qualcomm® SA7255P | QNN | 245.901 ms | 2 - 10 MB | NPU | Use Export Script |
|
| 47 |
+
| Depth-Anything-V2 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 147.49 ms | 0 - 102 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 48 |
+
| Depth-Anything-V2 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN | 130.649 ms | 3 - 6 MB | NPU | Use Export Script |
|
| 49 |
+
| Depth-Anything-V2 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 183.019 ms | 1 - 597 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 50 |
+
| Depth-Anything-V2 | float | SA8295P ADP | Qualcomm® SA8295P | QNN | 166.375 ms | 0 - 17 MB | NPU | Use Export Script |
|
| 51 |
+
| Depth-Anything-V2 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 148.76 ms | 0 - 108 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 52 |
+
| Depth-Anything-V2 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN | 133.085 ms | 3 - 7 MB | NPU | Use Export Script |
|
| 53 |
+
| Depth-Anything-V2 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 143.826 ms | 1 - 597 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 54 |
+
| Depth-Anything-V2 | float | SA8775P ADP | Qualcomm® SA8775P | QNN | 130.744 ms | 1 - 11 MB | NPU | Use Export Script |
|
| 55 |
+
| Depth-Anything-V2 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 138.816 ms | 0 - 102 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 56 |
+
| Depth-Anything-V2 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN | 126.571 ms | 3 - 102 MB | NPU | Use Export Script |
|
| 57 |
+
| Depth-Anything-V2 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 146.429 ms | 2 - 275 MB | NPU | [Depth-Anything-V2.onnx](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.onnx) |
|
| 58 |
+
| Depth-Anything-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 110.456 ms | 0 - 591 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 59 |
+
| Depth-Anything-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN | 99.074 ms | 3 - 730 MB | NPU | Use Export Script |
|
| 60 |
+
| Depth-Anything-V2 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 110.067 ms | 0 - 797 MB | NPU | [Depth-Anything-V2.onnx](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.onnx) |
|
| 61 |
+
| Depth-Anything-V2 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 98.459 ms | 0 - 562 MB | NPU | [Depth-Anything-V2.tflite](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.tflite) |
|
| 62 |
+
| Depth-Anything-V2 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN | 88.257 ms | 3 - 785 MB | NPU | Use Export Script |
|
| 63 |
+
| Depth-Anything-V2 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 91.333 ms | 3 - 886 MB | NPU | [Depth-Anything-V2.onnx](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.onnx) |
|
| 64 |
+
| Depth-Anything-V2 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 122.245 ms | 3 - 3 MB | NPU | Use Export Script |
|
| 65 |
+
| Depth-Anything-V2 | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 143.495 ms | 68 - 68 MB | NPU | [Depth-Anything-V2.onnx](https://huggingface.co/qualcomm/Depth-Anything-V2/blob/main/Depth-Anything-V2.onnx) |
|
| 66 |
|
| 67 |
|
| 68 |
|
|
|
|
| 126 |
Depth-Anything-V2
|
| 127 |
Device : cs_8275 (ANDROID 14)
|
| 128 |
Runtime : TFLITE
|
| 129 |
+
Estimated inference time (ms) : 889.5
|
| 130 |
+
Estimated peak memory usage (MB): [1, 598]
|
| 131 |
Total # Ops : 646
|
| 132 |
Compute Unit(s) : npu (646 ops) gpu (0 ops) cpu (0 ops)
|
| 133 |
```
|
|
|
|
| 216 |
You can also run the demo on-device.
|
| 217 |
|
| 218 |
```bash
|
| 219 |
+
python -m qai_hub_models.models.depth_anything_v2.demo --eval-mode on-device
|
| 220 |
```
|
| 221 |
|
| 222 |
**NOTE**: If you want running in a Jupyter Notebook or Google Colab like
|
| 223 |
environment, please add the following to your cell (instead of the above).
|
| 224 |
```
|
| 225 |
+
%run -m qai_hub_models.models.depth_anything_v2.demo -- --eval-mode on-device
|
| 226 |
```
|
| 227 |
|
| 228 |
|