Hugging Face's logo

qualcomm
/
Real-ESRGAN-General-x4v3

Image-to-Image
PyTorch
TF Lite
ONNX
android
Model card Files Community
bhushans commited on
Commit
a31e473
·
verified ·
1 Parent(s): 02ed120

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +44 -26
README.md CHANGED
@@ -34,29 +34,31 @@ More details on model performance across various devices, can be found
34
 
35
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
36
  |---|---|---|---|---|---|---|---|---|
37
- | Real-ESRGAN-General-x4v3 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 7.189 ms | 9 - 12 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
38
- | Real-ESRGAN-General-x4v3 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 6.279 ms | 0 - 107 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.so](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.so) |
39
- | Real-ESRGAN-General-x4v3 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 6.79 ms | 7 - 21 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.onnx](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.onnx) |
40
- | Real-ESRGAN-General-x4v3 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 5.387 ms | 9 - 72 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
41
- | Real-ESRGAN-General-x4v3 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 4.592 ms | 0 - 19 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.so](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.so) |
42
- | Real-ESRGAN-General-x4v3 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 4.986 ms | 6 - 76 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.onnx](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.onnx) |
43
- | Real-ESRGAN-General-x4v3 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 4.149 ms | 0 - 29 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
44
- | Real-ESRGAN-General-x4v3 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 4.325 ms | 0 - 18 MB | FP16 | NPU | Use Export Script |
45
- | Real-ESRGAN-General-x4v3 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 4.615 ms | 7 - 34 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.onnx](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.onnx) |
46
- | Real-ESRGAN-General-x4v3 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 7.275 ms | 6 - 13 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
47
- | Real-ESRGAN-General-x4v3 | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 5.731 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
48
- | Real-ESRGAN-General-x4v3 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 7.495 ms | 9 - 13 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
49
- | Real-ESRGAN-General-x4v3 | SA8255 (Proxy) | SA8255P Proxy | QNN | 5.765 ms | 0 - 4 MB | FP16 | NPU | Use Export Script |
50
- | Real-ESRGAN-General-x4v3 | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 7.34 ms | 9 - 123 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
51
- | Real-ESRGAN-General-x4v3 | SA8775 (Proxy) | SA8775P Proxy | QNN | 5.776 ms | 0 - 2 MB | FP16 | NPU | Use Export Script |
52
- | Real-ESRGAN-General-x4v3 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 7.297 ms | 9 - 127 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
53
- | Real-ESRGAN-General-x4v3 | SA8650 (Proxy) | SA8650P Proxy | QNN | 5.777 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
54
- | Real-ESRGAN-General-x4v3 | SA8295P ADP | SA8295P | TFLITE | 13.044 ms | 9 - 39 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
55
- | Real-ESRGAN-General-x4v3 | SA8295P ADP | SA8295P | QNN | 10.256 ms | 0 - 6 MB | FP16 | NPU | Use Export Script |
56
- | Real-ESRGAN-General-x4v3 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 12.381 ms | 9 - 75 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
57
- | Real-ESRGAN-General-x4v3 | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 9.701 ms | 0 - 24 MB | FP16 | NPU | Use Export Script |
58
- | Real-ESRGAN-General-x4v3 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 6.201 ms | 0 - 0 MB | FP16 | NPU | Use Export Script |
59
- | Real-ESRGAN-General-x4v3 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 7.125 ms | 8 - 8 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.onnx](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.onnx) |
 
 
60
 
61
 
62
 
@@ -121,8 +123,8 @@ Profiling Results
121
  Real-ESRGAN-General-x4v3
122
  Device : Samsung Galaxy S23 (13)
123
  Runtime : TFLITE
124
- Estimated inference time (ms) : 7.2
125
- Estimated peak memory usage (MB): [9, 12]
126
  Total # Ops : 72
127
  Compute Unit(s) : NPU (69 ops) CPU (3 ops)
128
  ```
@@ -143,13 +145,29 @@ in memory using the `jit.trace` and then call the `submit_compile_job` API.
143
  import torch
144
 
145
  import qai_hub as hub
146
- from qai_hub_models.models.real_esrgan_general_x4v3 import
147
 
148
  # Load the model
 
149
 
150
  # Device
151
  device = hub.Device("Samsung Galaxy S23")
152
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
153
 
154
  ```
155
 
 
34
 
35
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
36
  |---|---|---|---|---|---|---|---|---|
37
+ | Real-ESRGAN-General-x4v3 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 7.374 ms | 0 - 10 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
38
+ | Real-ESRGAN-General-x4v3 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 7.042 ms | 0 - 16 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.so](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.so) |
39
+ | Real-ESRGAN-General-x4v3 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 6.918 ms | 6 - 8 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.onnx](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.onnx) |
40
+ | Real-ESRGAN-General-x4v3 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 5.356 ms | 9 - 29 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
41
+ | Real-ESRGAN-General-x4v3 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 4.918 ms | 0 - 20 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.so](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.so) |
42
+ | Real-ESRGAN-General-x4v3 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 4.99 ms | 6 - 78 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.onnx](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.onnx) |
43
+ | Real-ESRGAN-General-x4v3 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 5.184 ms | 9 - 28 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
44
+ | Real-ESRGAN-General-x4v3 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 4.682 ms | 0 - 19 MB | FP16 | NPU | Use Export Script |
45
+ | Real-ESRGAN-General-x4v3 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 3.902 ms | 0 - 29 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.onnx](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.onnx) |
46
+ | Real-ESRGAN-General-x4v3 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 7.27 ms | 9 - 19 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
47
+ | Real-ESRGAN-General-x4v3 | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 6.192 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
48
+ | Real-ESRGAN-General-x4v3 | SA7255P ADP | SA7255P | TFLITE | 243.988 ms | 9 - 28 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
49
+ | Real-ESRGAN-General-x4v3 | SA7255P ADP | SA7255P | QNN | 237.989 ms | 0 - 6 MB | FP16 | NPU | Use Export Script |
50
+ | Real-ESRGAN-General-x4v3 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 7.358 ms | 9 - 20 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
51
+ | Real-ESRGAN-General-x4v3 | SA8255 (Proxy) | SA8255P Proxy | QNN | 6.212 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
52
+ | Real-ESRGAN-General-x4v3 | SA8295P ADP | SA8295P | TFLITE | 12.272 ms | 9 - 29 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
53
+ | Real-ESRGAN-General-x4v3 | SA8295P ADP | SA8295P | QNN | 11.334 ms | 0 - 6 MB | FP16 | NPU | Use Export Script |
54
+ | Real-ESRGAN-General-x4v3 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 7.493 ms | 9 - 19 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
55
+ | Real-ESRGAN-General-x4v3 | SA8650 (Proxy) | SA8650P Proxy | QNN | 6.227 ms | 0 - 2 MB | FP16 | NPU | Use Export Script |
56
+ | Real-ESRGAN-General-x4v3 | SA8775P ADP | SA8775P | TFLITE | 13.904 ms | 9 - 28 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
57
+ | Real-ESRGAN-General-x4v3 | SA8775P ADP | SA8775P | QNN | 12.075 ms | 0 - 6 MB | FP16 | NPU | Use Export Script |
58
+ | Real-ESRGAN-General-x4v3 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 12.394 ms | 9 - 37 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.tflite](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.tflite) |
59
+ | Real-ESRGAN-General-x4v3 | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 9.843 ms | 0 - 26 MB | FP16 | NPU | Use Export Script |
60
+ | Real-ESRGAN-General-x4v3 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 6.752 ms | 0 - 0 MB | FP16 | NPU | Use Export Script |
61
+ | Real-ESRGAN-General-x4v3 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 7.792 ms | 9 - 9 MB | FP16 | NPU | [Real-ESRGAN-General-x4v3.onnx](https://huggingface.co/qualcomm/Real-ESRGAN-General-x4v3/blob/main/Real-ESRGAN-General-x4v3.onnx) |
62
 
63
 
64
 
 
123
  Real-ESRGAN-General-x4v3
124
  Device : Samsung Galaxy S23 (13)
125
  Runtime : TFLITE
126
+ Estimated inference time (ms) : 7.4
127
+ Estimated peak memory usage (MB): [0, 10]
128
  Total # Ops : 72
129
  Compute Unit(s) : NPU (69 ops) CPU (3 ops)
130
  ```
 
145
  import torch
146
 
147
  import qai_hub as hub
148
+ from qai_hub_models.models.real_esrgan_general_x4v3 import Model
149
 
150
  # Load the model
151
+ torch_model = Model.from_pretrained()
152
 
153
  # Device
154
  device = hub.Device("Samsung Galaxy S23")
155
 
156
+ # Trace model
157
+ input_shape = torch_model.get_input_spec()
158
+ sample_inputs = torch_model.sample_inputs()
159
+
160
+ pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
161
+
162
+ # Compile model on a specific device
163
+ compile_job = hub.submit_compile_job(
164
+ model=pt_model,
165
+ device=device,
166
+ input_specs=torch_model.get_input_spec(),
167
+ )
168
+
169
+ # Get target model to run on-device
170
+ target_model = compile_job.get_target_model()
171
 
172
  ```
173