Hugging Face's logo

qualcomm
/
DDRNet23-Slim

Image Segmentation
PyTorch
TF Lite
ONNX
real_time
android
Model card Files Community
qaihm-bot commited on
Commit
9756ebc
•
1 Parent(s): c5ed1b1

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +32 -15
README.md CHANGED
@@ -37,19 +37,20 @@ More details on model performance across various devices, can be found
37
 
38
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
39
  |---|---|---|---|---|---|---|---|---|
40
- | DDRNet23-Slim | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 5.131 ms | 2 - 4 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
41
- | DDRNet23-Slim | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 7.388 ms | 9 - 14 MB | FP16 | NPU | [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) |
42
- | DDRNet23-Slim | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 3.489 ms | 1 - 77 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
43
- | DDRNet23-Slim | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 4.952 ms | 11 - 101 MB | FP16 | NPU | [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) |
44
- | DDRNet23-Slim | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 3.453 ms | 0 - 39 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
45
- | DDRNet23-Slim | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 4.99 ms | 11 - 55 MB | FP16 | NPU | [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) |
46
- | DDRNet23-Slim | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 5.068 ms | 0 - 2 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
47
- | DDRNet23-Slim | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 5.131 ms | 1 - 2 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
48
- | DDRNet23-Slim | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 5.185 ms | 0 - 2 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
49
- | DDRNet23-Slim | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 5.131 ms | 0 - 3 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
50
- | DDRNet23-Slim | SA8295P ADP | SA8295P | TFLITE | 8.857 ms | 1 - 31 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
51
- | DDRNet23-Slim | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 7.549 ms | 2 - 65 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
52
- | DDRNet23-Slim | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 8.362 ms | 9 - 9 MB | FP16 | NPU | [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) |
 
53
 
54
 
55
 
@@ -114,7 +115,7 @@ DDRNet23-Slim
114
  Device : Samsung Galaxy S23 (13)
115
  Runtime : TFLITE
116
  Estimated inference time (ms) : 5.1
117
- Estimated peak memory usage (MB): [2, 4]
118
  Total # Ops : 131
119
  Compute Unit(s) : NPU (131 ops)
120
  ```
@@ -135,13 +136,29 @@ in memory using the `jit.trace` and then call the `submit_compile_job` API.
135
  import torch
136
 
137
  import qai_hub as hub
138
- from qai_hub_models.models.ddrnet23_slim import
139
 
140
  # Load the model
 
141
 
142
  # Device
143
  device = hub.Device("Samsung Galaxy S23")
144
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
145
 
146
  ```
147
 
 
37
 
38
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
39
  |---|---|---|---|---|---|---|---|---|
40
+ | DDRNet23-Slim | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 5.087 ms | 1 - 20 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
41
+ | DDRNet23-Slim | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 7.462 ms | 9 - 25 MB | FP16 | NPU | [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) |
42
+ | DDRNet23-Slim | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 3.548 ms | 0 - 31 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
43
+ | DDRNet23-Slim | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 4.952 ms | 11 - 103 MB | FP16 | NPU | [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) |
44
+ | DDRNet23-Slim | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 2.854 ms | 0 - 28 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
45
+ | DDRNet23-Slim | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 4.11 ms | 11 - 58 MB | FP16 | NPU | [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) |
46
+ | DDRNet23-Slim | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 5.226 ms | 1 - 17 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
47
+ | DDRNet23-Slim | SA7255P ADP | SA7255P | TFLITE | 180.307 ms | 1 - 27 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
48
+ | DDRNet23-Slim | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 5.107 ms | 1 - 19 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
49
+ | DDRNet23-Slim | SA8295P ADP | SA8295P | TFLITE | 8.933 ms | 1 - 22 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
50
+ | DDRNet23-Slim | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 5.12 ms | 1 - 20 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
51
+ | DDRNet23-Slim | SA8775P ADP | SA8775P | TFLITE | 10.093 ms | 1 - 27 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
52
+ | DDRNet23-Slim | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 7.54 ms | 1 - 26 MB | FP16 | NPU | [DDRNet23-Slim.tflite](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.tflite) |
53
+ | DDRNet23-Slim | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 8.348 ms | 9 - 9 MB | FP16 | NPU | [DDRNet23-Slim.onnx](https://huggingface.co/qualcomm/DDRNet23-Slim/blob/main/DDRNet23-Slim.onnx) |
54
 
55
 
56
 
 
115
  Device : Samsung Galaxy S23 (13)
116
  Runtime : TFLITE
117
  Estimated inference time (ms) : 5.1
118
+ Estimated peak memory usage (MB): [1, 20]
119
  Total # Ops : 131
120
  Compute Unit(s) : NPU (131 ops)
121
  ```
 
136
  import torch
137
 
138
  import qai_hub as hub
139
+ from qai_hub_models.models.ddrnet23_slim import Model
140
 
141
  # Load the model
142
+ torch_model = Model.from_pretrained()
143
 
144
  # Device
145
  device = hub.Device("Samsung Galaxy S23")
146
 
147
+ # Trace model
148
+ input_shape = torch_model.get_input_spec()
149
+ sample_inputs = torch_model.sample_inputs()
150
+
151
+ pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])
152
+
153
+ # Compile model on a specific device
154
+ compile_job = hub.submit_compile_job(
155
+ model=pt_model,
156
+ device=device,
157
+ input_specs=torch_model.get_input_spec(),
158
+ )
159
+
160
+ # Get target model to run on-device
161
+ target_model = compile_job.get_target_model()
162
 
163
  ```
164