Hugging Face's logo

qualcomm
/
DETR-ResNet50

Object Detection
PyTorch
LiteRT
bu_auto
android
Model card Files
xet
Community
1
qaihm-bot commited on
Commit
2fe4b97
·
verified ·
1 Parent(s): be22515

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +32 -16
README.md CHANGED
@@ -34,19 +34,19 @@ 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
- | DETR-ResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 11.032 ms | 0 - 2 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
38
- | DETR-ResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 12.244 ms | 0 - 97 MB | FP16 | NPU | [DETR-ResNet50.onnx](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.onnx) |
39
- | DETR-ResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 7.656 ms | 0 - 246 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
40
- | DETR-ResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 8.821 ms | 3 - 215 MB | FP16 | NPU | [DETR-ResNet50.onnx](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.onnx) |
41
- | DETR-ResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 7.217 ms | 0 - 89 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
42
- | DETR-ResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 8.47 ms | 3 - 90 MB | FP16 | NPU | [DETR-ResNet50.onnx](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.onnx) |
43
- | DETR-ResNet50 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 10.807 ms | 0 - 2 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
44
- | DETR-ResNet50 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 10.957 ms | 0 - 2 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
45
- | DETR-ResNet50 | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 10.897 ms | 0 - 2 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
46
- | DETR-ResNet50 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 10.941 ms | 0 - 2 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
47
- | DETR-ResNet50 | SA8295P ADP | SA8295P | TFLITE | 16.44 ms | 0 - 64 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
48
- | DETR-ResNet50 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 14.43 ms | 0 - 212 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
49
- | DETR-ResNet50 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 13.375 ms | 80 - 80 MB | FP16 | NPU | [DETR-ResNet50.onnx](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.onnx) |
50
 
51
 
52
 
@@ -111,8 +111,8 @@ Profiling Results
111
  DETR-ResNet50
112
  Device : Samsung Galaxy S23 (13)
113
  Runtime : TFLITE
114
- Estimated inference time (ms) : 11.0
115
- Estimated peak memory usage (MB): [0, 2]
116
  Total # Ops : 788
117
  Compute Unit(s) : NPU (788 ops)
118
  ```
@@ -133,13 +133,29 @@ in memory using the `jit.trace` and then call the `submit_compile_job` API.
133
  import torch
134
 
135
  import qai_hub as hub
136
- from qai_hub_models.models.detr_resnet50 import
137
 
138
  # Load the model
 
139
 
140
  # Device
141
  device = hub.Device("Samsung Galaxy S23")
142
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
143
 
144
  ```
145
 
 
34
 
35
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
36
  |---|---|---|---|---|---|---|---|---|
37
+ | DETR-ResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 10.819 ms | 0 - 24 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
38
+ | DETR-ResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 12.164 ms | 0 - 406 MB | FP16 | NPU | [DETR-ResNet50.onnx](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.onnx) |
39
+ | DETR-ResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 7.712 ms | 0 - 66 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
40
+ | DETR-ResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 8.834 ms | 3 - 217 MB | FP16 | NPU | [DETR-ResNet50.onnx](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.onnx) |
41
+ | DETR-ResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 7.269 ms | 0 - 65 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
42
+ | DETR-ResNet50 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 10.852 ms | 0 - 26 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
43
+ | DETR-ResNet50 | SA7255P ADP | SA7255P | TFLITE | 309.163 ms | 0 - 65 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
44
+ | DETR-ResNet50 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 10.817 ms | 0 - 26 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
45
+ | DETR-ResNet50 | SA8295P ADP | SA8295P | TFLITE | 16.397 ms | 0 - 46 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
46
+ | DETR-ResNet50 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 10.831 ms | 0 - 32 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
47
+ | DETR-ResNet50 | SA8775P ADP | SA8775P | TFLITE | 18.402 ms | 0 - 64 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
48
+ | DETR-ResNet50 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 14.437 ms | 0 - 47 MB | FP16 | NPU | [DETR-ResNet50.tflite](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.tflite) |
49
+ | DETR-ResNet50 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 13.381 ms | 79 - 79 MB | FP16 | NPU | [DETR-ResNet50.onnx](https://huggingface.co/qualcomm/DETR-ResNet50/blob/main/DETR-ResNet50.onnx) |
50
 
51
 
52
 
 
111
  DETR-ResNet50
112
  Device : Samsung Galaxy S23 (13)
113
  Runtime : TFLITE
114
+ Estimated inference time (ms) : 10.8
115
+ Estimated peak memory usage (MB): [0, 24]
116
  Total # Ops : 788
117
  Compute Unit(s) : NPU (788 ops)
118
  ```
 
133
  import torch
134
 
135
  import qai_hub as hub
136
+ from qai_hub_models.models.detr_resnet50 import Model
137
 
138
  # Load the model
139
+ torch_model = Model.from_pretrained()
140
 
141
  # Device
142
  device = hub.Device("Samsung Galaxy S23")
143
 
144
+ # Trace model
145
+ input_shape = torch_model.get_input_spec()
146
+ sample_inputs = torch_model.sample_inputs()
147
+
148
+ pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
149
+
150
+ # Compile model on a specific device
151
+ compile_job = hub.submit_compile_job(
152
+ model=pt_model,
153
+ device=device,
154
+ input_specs=torch_model.get_input_spec(),
155
+ )
156
+
157
+ # Get target model to run on-device
158
+ target_model = compile_job.get_target_model()
159
 
160
  ```
161