Hugging Face's logo

qualcomm
/
FFNet-40S

Image Segmentation
PyTorch
LiteRT
real_time
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
25280c2
·
verified ·
1 Parent(s): e8b722f

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +32 -32
README.md CHANGED
@@ -36,45 +36,44 @@ More details on model performance across various devices, can be found
36
 
37
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
- | FFNet-40S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 45.422 ms | 4 - 21 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
40
- | FFNet-40S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 54.991 ms | 24 - 41 MB | FP16 | NPU | [FFNet-40S.so](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.so) |
41
- | FFNet-40S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 35.72 ms | 29 - 101 MB | FP16 | NPU | [FFNet-40S.onnx](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.onnx) |
42
- | FFNet-40S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 29.716 ms | 2 - 49 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
43
- | FFNet-40S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 37.476 ms | 14 - 56 MB | FP16 | NPU | [FFNet-40S.so](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.so) |
44
- | FFNet-40S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 25.75 ms | 26 - 70 MB | FP16 | NPU | [FFNet-40S.onnx](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.onnx) |
45
- | FFNet-40S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 23.928 ms | 1 - 49 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
46
- | FFNet-40S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 32.713 ms | 12 - 54 MB | FP16 | NPU | Use Export Script |
47
- | FFNet-40S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 24.234 ms | 30 - 72 MB | FP16 | NPU | [FFNet-40S.onnx](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.onnx) |
48
- | FFNet-40S | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 44.925 ms | 2 - 22 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
49
- | FFNet-40S | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 41.513 ms | 24 - 27 MB | FP16 | NPU | Use Export Script |
50
- | FFNet-40S | SA7255P ADP | SA7255P | TFLITE | 853.906 ms | 3 - 46 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
51
- | FFNet-40S | SA7255P ADP | SA7255P | QNN | 842.941 ms | 24 - 34 MB | FP16 | NPU | Use Export Script |
52
- | FFNet-40S | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 45.165 ms | 4 - 26 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
53
- | FFNet-40S | SA8255 (Proxy) | SA8255P Proxy | QNN | 41.508 ms | 24 - 26 MB | FP16 | NPU | Use Export Script |
54
- | FFNet-40S | SA8295P ADP | SA8295P | TFLITE | 66.395 ms | 2 - 45 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
55
- | FFNet-40S | SA8295P ADP | SA8295P | QNN | 64.583 ms | 24 - 38 MB | FP16 | NPU | Use Export Script |
56
- | FFNet-40S | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 45.194 ms | 2 - 23 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
57
- | FFNet-40S | SA8650 (Proxy) | SA8650P Proxy | QNN | 41.421 ms | 24 - 26 MB | FP16 | NPU | Use Export Script |
58
- | FFNet-40S | SA8775P ADP | SA8775P | TFLITE | 71.816 ms | 2 - 46 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
59
- | FFNet-40S | SA8775P ADP | SA8775P | QNN | 65.545 ms | 24 - 34 MB | FP16 | NPU | Use Export Script |
60
- | FFNet-40S | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 57.324 ms | 2 - 47 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
61
- | FFNet-40S | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 69.305 ms | 23 - 62 MB | FP16 | NPU | Use Export Script |
62
- | FFNet-40S | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 42.428 ms | 24 - 24 MB | FP16 | NPU | Use Export Script |
63
- | FFNet-40S | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 38.682 ms | 24 - 24 MB | FP16 | NPU | [FFNet-40S.onnx](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.onnx) |
64
 
65
 
66
 
67
 
68
  ## Installation
69
 
70
- This model can be installed as a Python package via pip.
71
 
 
72
  ```bash
73
- pip install "qai-hub-models[ffnet_40s]"
74
  ```
75
 
76
 
77
-
78
  ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
79
 
80
  Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
@@ -125,8 +124,8 @@ Profiling Results
125
  FFNet-40S
126
  Device : Samsung Galaxy S23 (13)
127
  Runtime : TFLITE
128
- Estimated inference time (ms) : 45.4
129
- Estimated peak memory usage (MB): [4, 21]
130
  Total # Ops : 94
131
  Compute Unit(s) : NPU (94 ops)
132
  ```
@@ -153,7 +152,7 @@ from qai_hub_models.models.ffnet_40s import Model
153
  torch_model = Model.from_pretrained()
154
 
155
  # Device
156
- device = hub.Device("Samsung Galaxy S23")
157
 
158
  # Trace model
159
  input_shape = torch_model.get_input_spec()
@@ -245,7 +244,8 @@ Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
245
 
246
 
247
  ## License
248
- * The license for the original implementation of FFNet-40S can be found [here](https://github.com/Qualcomm-AI-research/FFNet/blob/master/LICENSE).
 
249
  * The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf)
250
 
251
 
 
36
 
37
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
+ | FFNet-40S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 45.018 ms | 2 - 26 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
40
+ | FFNet-40S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 55.016 ms | 24 - 41 MB | FP16 | NPU | [FFNet-40S.so](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.so) |
41
+ | FFNet-40S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 34.72 ms | 29 - 119 MB | FP16 | NPU | [FFNet-40S.onnx](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.onnx) |
42
+ | FFNet-40S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 29.645 ms | 1 - 50 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
43
+ | FFNet-40S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 37.001 ms | 9 - 49 MB | FP16 | NPU | [FFNet-40S.so](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.so) |
44
+ | FFNet-40S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 26.015 ms | 30 - 76 MB | FP16 | NPU | [FFNet-40S.onnx](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.onnx) |
45
+ | FFNet-40S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 29.038 ms | 9 - 58 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
46
+ | FFNet-40S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 33.165 ms | 24 - 67 MB | FP16 | NPU | Use Export Script |
47
+ | FFNet-40S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 23.801 ms | 28 - 70 MB | FP16 | NPU | [FFNet-40S.onnx](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.onnx) |
48
+ | FFNet-40S | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 44.885 ms | 2 - 24 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
49
+ | FFNet-40S | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 41.396 ms | 24 - 27 MB | FP16 | NPU | Use Export Script |
50
+ | FFNet-40S | SA7255P ADP | SA7255P | TFLITE | 854.012 ms | 2 - 45 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
51
+ | FFNet-40S | SA7255P ADP | SA7255P | QNN | 842.577 ms | 21 - 31 MB | FP16 | NPU | Use Export Script |
52
+ | FFNet-40S | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 45.077 ms | 2 - 21 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
53
+ | FFNet-40S | SA8255 (Proxy) | SA8255P Proxy | QNN | 41.467 ms | 24 - 26 MB | FP16 | NPU | Use Export Script |
54
+ | FFNet-40S | SA8295P ADP | SA8295P | TFLITE | 66.248 ms | 2 - 45 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
55
+ | FFNet-40S | SA8295P ADP | SA8295P | QNN | 64.495 ms | 24 - 38 MB | FP16 | NPU | Use Export Script |
56
+ | FFNet-40S | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 45.311 ms | 2 - 24 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
57
+ | FFNet-40S | SA8650 (Proxy) | SA8650P Proxy | QNN | 41.258 ms | 24 - 27 MB | FP16 | NPU | Use Export Script |
58
+ | FFNet-40S | SA8775P ADP | SA8775P | TFLITE | 71.761 ms | 2 - 45 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
59
+ | FFNet-40S | SA8775P ADP | SA8775P | QNN | 65.652 ms | 24 - 34 MB | FP16 | NPU | Use Export Script |
60
+ | FFNet-40S | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 57.971 ms | 2 - 49 MB | FP16 | NPU | [FFNet-40S.tflite](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.tflite) |
61
+ | FFNet-40S | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 69.253 ms | 23 - 60 MB | FP16 | NPU | Use Export Script |
62
+ | FFNet-40S | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 42.478 ms | 24 - 24 MB | FP16 | NPU | Use Export Script |
63
+ | FFNet-40S | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 38.729 ms | 24 - 24 MB | FP16 | NPU | [FFNet-40S.onnx](https://huggingface.co/qualcomm/FFNet-40S/blob/main/FFNet-40S.onnx) |
64
 
65
 
66
 
67
 
68
  ## Installation
69
 
 
70
 
71
+ Install the package via pip:
72
  ```bash
73
+ pip install "qai-hub-models[ffnet-40s]"
74
  ```
75
 
76
 
 
77
  ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
78
 
79
  Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
 
124
  FFNet-40S
125
  Device : Samsung Galaxy S23 (13)
126
  Runtime : TFLITE
127
+ Estimated inference time (ms) : 45.0
128
+ Estimated peak memory usage (MB): [2, 26]
129
  Total # Ops : 94
130
  Compute Unit(s) : NPU (94 ops)
131
  ```
 
152
  torch_model = Model.from_pretrained()
153
 
154
  # Device
155
+ device = hub.Device("Samsung Galaxy S24")
156
 
157
  # Trace model
158
  input_shape = torch_model.get_input_spec()
 
244
 
245
 
246
  ## License
247
+ * The license for the original implementation of FFNet-40S can be found
248
+ [here](https://github.com/Qualcomm-AI-research/FFNet/blob/master/LICENSE).
249
  * The license for the compiled assets for on-device deployment can be found [here](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/Qualcomm+AI+Hub+Proprietary+License.pdf)
250
 
251