Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -38,50 +38,54 @@ More details on model performance across various devices, can be found
|
|
| 38 |
|
| 39 |
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
|
| 40 |
|---|---|---|---|---|---|---|---|---|
|
| 41 |
-
| CLIPTextEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 5.
|
| 42 |
-
| CLIPTextEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 4.
|
| 43 |
-
| CLIPTextEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX |
|
| 44 |
-
| CLIPTextEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE |
|
| 45 |
-
| CLIPTextEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 3.
|
| 46 |
-
| CLIPTextEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX |
|
| 47 |
-
| CLIPTextEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 3.
|
| 48 |
-
| CLIPTextEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN |
|
| 49 |
-
| CLIPTextEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX |
|
| 50 |
-
| CLIPTextEncoder | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 5.
|
| 51 |
-
| CLIPTextEncoder | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 4.
|
| 52 |
-
| CLIPTextEncoder |
|
| 53 |
-
| CLIPTextEncoder |
|
| 54 |
-
| CLIPTextEncoder |
|
| 55 |
-
| CLIPTextEncoder |
|
| 56 |
-
| CLIPTextEncoder |
|
| 57 |
-
| CLIPTextEncoder |
|
| 58 |
-
| CLIPTextEncoder |
|
| 59 |
-
| CLIPTextEncoder |
|
| 60 |
-
| CLIPTextEncoder |
|
| 61 |
-
| CLIPTextEncoder |
|
| 62 |
-
| CLIPTextEncoder |
|
| 63 |
-
| CLIPTextEncoder |
|
| 64 |
-
|
|
| 65 |
-
|
|
| 66 |
-
| CLIPImageEncoder | Samsung Galaxy
|
| 67 |
-
| CLIPImageEncoder | Samsung Galaxy
|
| 68 |
-
| CLIPImageEncoder | Samsung Galaxy
|
| 69 |
-
| CLIPImageEncoder |
|
| 70 |
-
| CLIPImageEncoder |
|
| 71 |
-
| CLIPImageEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite |
|
| 72 |
-
| CLIPImageEncoder |
|
| 73 |
-
| CLIPImageEncoder |
|
| 74 |
-
| CLIPImageEncoder |
|
| 75 |
-
| CLIPImageEncoder |
|
| 76 |
-
| CLIPImageEncoder |
|
| 77 |
-
| CLIPImageEncoder |
|
| 78 |
-
| CLIPImageEncoder |
|
| 79 |
-
| CLIPImageEncoder |
|
| 80 |
| CLIPImageEncoder | SA8295P ADP | SA8295P | TFLITE | 42.408 ms | 0 - 359 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 81 |
-
| CLIPImageEncoder | SA8295P ADP | SA8295P | QNN |
|
| 82 |
-
| CLIPImageEncoder |
|
| 83 |
-
| CLIPImageEncoder |
|
| 84 |
-
| CLIPImageEncoder |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 85 |
|
| 86 |
|
| 87 |
|
|
@@ -147,7 +151,7 @@ CLIPTextEncoder
|
|
| 147 |
Device : Samsung Galaxy S23 (13)
|
| 148 |
Runtime : TFLITE
|
| 149 |
Estimated inference time (ms) : 5.7
|
| 150 |
-
Estimated peak memory usage (MB): [0,
|
| 151 |
Total # Ops : 660
|
| 152 |
Compute Unit(s) : NPU (658 ops) CPU (2 ops)
|
| 153 |
|
|
@@ -155,8 +159,8 @@ Compute Unit(s) : NPU (658 ops) CPU (2 ops)
|
|
| 155 |
CLIPImageEncoder
|
| 156 |
Device : Samsung Galaxy S23 (13)
|
| 157 |
Runtime : TFLITE
|
| 158 |
-
Estimated inference time (ms) :
|
| 159 |
-
Estimated peak memory usage (MB): [0,
|
| 160 |
Total # Ops : 659
|
| 161 |
Compute Unit(s) : NPU (659 ops)
|
| 162 |
```
|
|
@@ -177,11 +181,12 @@ in memory using the `jit.trace` and then call the `submit_compile_job` API.
|
|
| 177 |
import torch
|
| 178 |
|
| 179 |
import qai_hub as hub
|
| 180 |
-
from qai_hub_models.models.openai_clip import
|
| 181 |
|
| 182 |
# Load the model
|
| 183 |
-
|
| 184 |
-
|
|
|
|
| 185 |
|
| 186 |
# Device
|
| 187 |
device = hub.Device("Samsung Galaxy S23")
|
|
|
|
| 38 |
|
| 39 |
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
|
| 40 |
|---|---|---|---|---|---|---|---|---|
|
| 41 |
+
| CLIPTextEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 5.678 ms | 0 - 17 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 42 |
+
| CLIPTextEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 4.678 ms | 0 - 18 MB | FP16 | NPU | [OpenAI-Clip.so](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.so) |
|
| 43 |
+
| CLIPTextEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 34.783 ms | 0 - 1424 MB | FP16 | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.onnx) |
|
| 44 |
+
| CLIPTextEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 3.997 ms | 0 - 84 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 45 |
+
| CLIPTextEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 3.288 ms | 0 - 70 MB | FP16 | NPU | [OpenAI-Clip.so](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.so) |
|
| 46 |
+
| CLIPTextEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 24.961 ms | 0 - 531 MB | FP16 | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.onnx) |
|
| 47 |
+
| CLIPTextEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 3.961 ms | 0 - 82 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 48 |
+
| CLIPTextEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 3.27 ms | 0 - 67 MB | FP16 | NPU | Use Export Script |
|
| 49 |
+
| CLIPTextEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 20.656 ms | 0 - 316 MB | FP16 | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.onnx) |
|
| 50 |
+
| CLIPTextEncoder | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 5.589 ms | 0 - 18 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 51 |
+
| CLIPTextEncoder | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 4.685 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
|
| 52 |
+
| CLIPTextEncoder | SA7255P ADP | SA7255P | TFLITE | 61.394 ms | 0 - 81 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 53 |
+
| CLIPTextEncoder | SA7255P ADP | SA7255P | QNN | 51.693 ms | 0 - 6 MB | FP16 | NPU | Use Export Script |
|
| 54 |
+
| CLIPTextEncoder | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 5.678 ms | 0 - 18 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 55 |
+
| CLIPTextEncoder | SA8255 (Proxy) | SA8255P Proxy | QNN | 4.766 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
|
| 56 |
+
| CLIPTextEncoder | SA8295P ADP | SA8295P | TFLITE | 7.639 ms | 0 - 67 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 57 |
+
| CLIPTextEncoder | SA8295P ADP | SA8295P | QNN | 6.535 ms | 0 - 6 MB | FP16 | NPU | Use Export Script |
|
| 58 |
+
| CLIPTextEncoder | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 5.695 ms | 0 - 18 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 59 |
+
| CLIPTextEncoder | SA8650 (Proxy) | SA8650P Proxy | QNN | 4.767 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
|
| 60 |
+
| CLIPTextEncoder | SA8775P ADP | SA8775P | TFLITE | 8.155 ms | 0 - 80 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 61 |
+
| CLIPTextEncoder | SA8775P ADP | SA8775P | QNN | 6.942 ms | 0 - 5 MB | FP16 | NPU | Use Export Script |
|
| 62 |
+
| CLIPTextEncoder | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 6.336 ms | 0 - 69 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.tflite) |
|
| 63 |
+
| CLIPTextEncoder | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 5.217 ms | 0 - 69 MB | FP16 | NPU | Use Export Script |
|
| 64 |
+
| CLIPTextEncoder | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 5.156 ms | 0 - 0 MB | FP16 | NPU | Use Export Script |
|
| 65 |
+
| CLIPTextEncoder | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 38.009 ms | 126 - 126 MB | FP16 | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPTextEncoder.onnx) |
|
| 66 |
+
| CLIPImageEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 34.201 ms | 0 - 51 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 67 |
+
| CLIPImageEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 26.406 ms | 0 - 54 MB | FP16 | NPU | [OpenAI-Clip.so](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.so) |
|
| 68 |
+
| CLIPImageEncoder | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 158.287 ms | 0 - 194 MB | FP16 | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.onnx) |
|
| 69 |
+
| CLIPImageEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 26.701 ms | 0 - 262 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 70 |
+
| CLIPImageEncoder | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 20.878 ms | 49 - 220 MB | FP16 | NPU | [OpenAI-Clip.so](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.so) |
|
| 71 |
+
| CLIPImageEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 21.369 ms | 0 - 265 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 72 |
+
| CLIPImageEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 18.876 ms | 0 - 172 MB | FP16 | NPU | Use Export Script |
|
| 73 |
+
| CLIPImageEncoder | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 95.268 ms | 1 - 2718 MB | FP16 | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.onnx) |
|
| 74 |
+
| CLIPImageEncoder | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 33.57 ms | 0 - 52 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 75 |
+
| CLIPImageEncoder | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 20.122 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
|
| 76 |
+
| CLIPImageEncoder | SA7255P ADP | SA7255P | TFLITE | 326.507 ms | 0 - 264 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 77 |
+
| CLIPImageEncoder | SA7255P ADP | SA7255P | QNN | 265.126 ms | 1 - 7 MB | FP16 | NPU | Use Export Script |
|
| 78 |
+
| CLIPImageEncoder | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 34.116 ms | 0 - 56 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 79 |
+
| CLIPImageEncoder | SA8255 (Proxy) | SA8255P Proxy | QNN | 20.571 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
|
| 80 |
| CLIPImageEncoder | SA8295P ADP | SA8295P | TFLITE | 42.408 ms | 0 - 359 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 81 |
+
| CLIPImageEncoder | SA8295P ADP | SA8295P | QNN | 30.852 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
|
| 82 |
+
| CLIPImageEncoder | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 34.117 ms | 0 - 57 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 83 |
+
| CLIPImageEncoder | SA8650 (Proxy) | SA8650P Proxy | QNN | 20.427 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
|
| 84 |
+
| CLIPImageEncoder | SA8775P ADP | SA8775P | QNN | 29.742 ms | 0 - 5 MB | FP16 | NPU | Use Export Script |
|
| 85 |
+
| CLIPImageEncoder | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 34.821 ms | 0 - 203 MB | FP16 | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.tflite) |
|
| 86 |
+
| CLIPImageEncoder | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 29.464 ms | 0 - 169 MB | FP16 | NPU | Use Export Script |
|
| 87 |
+
| CLIPImageEncoder | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 22.2 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
|
| 88 |
+
| CLIPImageEncoder | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 160.456 ms | 188 - 188 MB | FP16 | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/CLIPImageEncoder.onnx) |
|
| 89 |
|
| 90 |
|
| 91 |
|
|
|
|
| 151 |
Device : Samsung Galaxy S23 (13)
|
| 152 |
Runtime : TFLITE
|
| 153 |
Estimated inference time (ms) : 5.7
|
| 154 |
+
Estimated peak memory usage (MB): [0, 17]
|
| 155 |
Total # Ops : 660
|
| 156 |
Compute Unit(s) : NPU (658 ops) CPU (2 ops)
|
| 157 |
|
|
|
|
| 159 |
CLIPImageEncoder
|
| 160 |
Device : Samsung Galaxy S23 (13)
|
| 161 |
Runtime : TFLITE
|
| 162 |
+
Estimated inference time (ms) : 34.2
|
| 163 |
+
Estimated peak memory usage (MB): [0, 51]
|
| 164 |
Total # Ops : 659
|
| 165 |
Compute Unit(s) : NPU (659 ops)
|
| 166 |
```
|
|
|
|
| 181 |
import torch
|
| 182 |
|
| 183 |
import qai_hub as hub
|
| 184 |
+
from qai_hub_models.models.openai_clip import Model
|
| 185 |
|
| 186 |
# Load the model
|
| 187 |
+
model = Model.from_pretrained()
|
| 188 |
+
text_encoder_model = model.text_encoder
|
| 189 |
+
image_encoder_model = model.image_encoder
|
| 190 |
|
| 191 |
# Device
|
| 192 |
device = hub.Device("Samsung Galaxy S23")
|