Hugging Face's logo

qualcomm
/
Midas-V2

Depth Estimation
PyTorch
LiteRT
ONNX
android
Model card Files Community
bhushans commited on
Commit
bb5d8c8
·
verified ·
1 Parent(s): 5f8b9eb

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +43 -25
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
- | Midas-V2 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 3.227 ms | 0 - 9 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
38
- | Midas-V2 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 3.276 ms | 0 - 111 MB | FP16 | NPU | [Midas-V2.so](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.so) |
39
- | Midas-V2 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 3.295 ms | 0 - 41 MB | FP16 | NPU | [Midas-V2.onnx](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.onnx) |
40
- | Midas-V2 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 2.27 ms | 0 - 89 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
41
- | Midas-V2 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 2.312 ms | 1 - 28 MB | FP16 | NPU | [Midas-V2.so](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.so) |
42
- | Midas-V2 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 2.398 ms | 0 - 92 MB | FP16 | NPU | [Midas-V2.onnx](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.onnx) |
43
- | Midas-V2 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 2.133 ms | 0 - 38 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
44
- | Midas-V2 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 2.16 ms | 0 - 23 MB | FP16 | NPU | Use Export Script |
45
- | Midas-V2 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 2.219 ms | 0 - 42 MB | FP16 | NPU | [Midas-V2.onnx](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.onnx) |
46
- | Midas-V2 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 3.217 ms | 0 - 2 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
47
- | Midas-V2 | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 3.022 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
48
- | Midas-V2 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 3.239 ms | 0 - 2 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
49
- | Midas-V2 | SA8255 (Proxy) | SA8255P Proxy | QNN | 3.04 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
50
- | Midas-V2 | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 3.227 ms | 0 - 2 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
51
- | Midas-V2 | SA8775 (Proxy) | SA8775P Proxy | QNN | 3.051 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
52
- | Midas-V2 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 3.236 ms | 0 - 2 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
53
- | Midas-V2 | SA8650 (Proxy) | SA8650P Proxy | QNN | 3.051 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
54
- | Midas-V2 | SA8295P ADP | SA8295P | TFLITE | 5.626 ms | 0 - 40 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
55
- | Midas-V2 | SA8295P ADP | SA8295P | QNN | 5.459 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
56
- | Midas-V2 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 4.734 ms | 0 - 92 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
57
- | Midas-V2 | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 4.919 ms | 1 - 25 MB | FP16 | NPU | Use Export Script |
58
- | Midas-V2 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 3.207 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
59
- | Midas-V2 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.354 ms | 36 - 36 MB | FP16 | NPU | [Midas-V2.onnx](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.onnx) |
 
 
60
 
61
 
62
 
@@ -122,7 +124,7 @@ Midas-V2
122
  Device : Samsung Galaxy S23 (13)
123
  Runtime : TFLITE
124
  Estimated inference time (ms) : 3.2
125
- Estimated peak memory usage (MB): [0, 9]
126
  Total # Ops : 138
127
  Compute Unit(s) : NPU (138 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.midas 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
+ | Midas-V2 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 3.249 ms | 0 - 19 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
38
+ | Midas-V2 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 3.275 ms | 0 - 111 MB | FP16 | NPU | [Midas-V2.so](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.so) |
39
+ | Midas-V2 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 3.304 ms | 0 - 41 MB | FP16 | NPU | [Midas-V2.onnx](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.onnx) |
40
+ | Midas-V2 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 2.269 ms | 0 - 27 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
41
+ | Midas-V2 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 2.308 ms | 1 - 25 MB | FP16 | NPU | [Midas-V2.so](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.so) |
42
+ | Midas-V2 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 2.539 ms | 0 - 93 MB | FP16 | NPU | [Midas-V2.onnx](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.onnx) |
43
+ | Midas-V2 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 1.78 ms | 0 - 23 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
44
+ | Midas-V2 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 2.17 ms | 0 - 22 MB | FP16 | NPU | Use Export Script |
45
+ | Midas-V2 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 2.226 ms | 0 - 43 MB | FP16 | NPU | [Midas-V2.onnx](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.onnx) |
46
+ | Midas-V2 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 3.213 ms | 0 - 51 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
47
+ | Midas-V2 | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 3.027 ms | 1 - 2 MB | FP16 | NPU | Use Export Script |
48
+ | Midas-V2 | SA7255P ADP | SA7255P | TFLITE | 84.525 ms | 0 - 22 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
49
+ | Midas-V2 | SA7255P ADP | SA7255P | QNN | 84.275 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
50
+ | Midas-V2 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 3.225 ms | 0 - 40 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
51
+ | Midas-V2 | SA8255 (Proxy) | SA8255P Proxy | QNN | 3.043 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
52
+ | Midas-V2 | SA8295P ADP | SA8295P | TFLITE | 5.59 ms | 0 - 21 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
53
+ | Midas-V2 | SA8295P ADP | SA8295P | QNN | 5.669 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
54
+ | Midas-V2 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 3.245 ms | 0 - 21 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
55
+ | Midas-V2 | SA8650 (Proxy) | SA8650P Proxy | QNN | 3.034 ms | 0 - 1 MB | FP16 | NPU | Use Export Script |
56
+ | Midas-V2 | SA8775P ADP | SA8775P | TFLITE | 5.435 ms | 0 - 21 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
57
+ | Midas-V2 | SA8775P ADP | SA8775P | QNN | 5.233 ms | 1 - 6 MB | FP16 | NPU | Use Export Script |
58
+ | Midas-V2 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 4.787 ms | 0 - 24 MB | FP16 | NPU | [Midas-V2.tflite](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.tflite) |
59
+ | Midas-V2 | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 4.906 ms | 1 - 23 MB | FP16 | NPU | Use Export Script |
60
+ | Midas-V2 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 3.23 ms | 1 - 1 MB | FP16 | NPU | Use Export Script |
61
+ | Midas-V2 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 3.344 ms | 37 - 37 MB | FP16 | NPU | [Midas-V2.onnx](https://huggingface.co/qualcomm/Midas-V2/blob/main/Midas-V2.onnx) |
62
 
63
 
64
 
 
124
  Device : Samsung Galaxy S23 (13)
125
  Runtime : TFLITE
126
  Estimated inference time (ms) : 3.2
127
+ Estimated peak memory usage (MB): [0, 19]
128
  Total # Ops : 138
129
  Compute Unit(s) : NPU (138 ops)
130
  ```
 
145
  import torch
146
 
147
  import qai_hub as hub
148
+ from qai_hub_models.models.midas 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