Hugging Face's logo

qualcomm
/
FFNet-78S

Image Segmentation
PyTorch
LiteRT
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
db5fb2d
·
verified ·
1 Parent(s): 3879c1e

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +32 -32
README.md CHANGED
@@ -35,45 +35,44 @@ More details on model performance across various devices, can be found
35
 
36
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
38
- | FFNet-78S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 50.728 ms | 2 - 34 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
39
- | FFNet-78S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 61.315 ms | 24 - 44 MB | FP16 | NPU | [FFNet-78S.so](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.so) |
40
- | FFNet-78S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 39.439 ms | 29 - 179 MB | FP16 | NPU | [FFNet-78S.onnx](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.onnx) |
41
- | FFNet-78S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 34.633 ms | 1 - 55 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
42
- | FFNet-78S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 41.943 ms | 24 - 71 MB | FP16 | NPU | [FFNet-78S.so](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.so) |
43
- | FFNet-78S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 32.743 ms | 28 - 90 MB | FP16 | NPU | [FFNet-78S.onnx](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.onnx) |
44
- | FFNet-78S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 33.453 ms | 1 - 54 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
45
- | FFNet-78S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 44.136 ms | 24 - 71 MB | FP16 | NPU | Use Export Script |
46
- | FFNet-78S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 29.256 ms | 29 - 82 MB | FP16 | NPU | [FFNet-78S.onnx](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.onnx) |
47
- | FFNet-78S | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 51.023 ms | 2 - 36 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
48
- | FFNet-78S | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 46.691 ms | 24 - 27 MB | FP16 | NPU | Use Export Script |
49
- | FFNet-78S | SA7255P ADP | SA7255P | TFLITE | 1255.866 ms | 0 - 51 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
50
- | FFNet-78S | SA7255P ADP | SA7255P | QNN | 1243.774 ms | 24 - 34 MB | FP16 | NPU | Use Export Script |
51
- | FFNet-78S | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 50.834 ms | 3 - 39 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
52
- | FFNet-78S | SA8255 (Proxy) | SA8255P Proxy | QNN | 46.909 ms | 24 - 27 MB | FP16 | NPU | Use Export Script |
53
- | FFNet-78S | SA8295P ADP | SA8295P | TFLITE | 78.938 ms | 2 - 49 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
54
- | FFNet-78S | SA8295P ADP | SA8295P | QNN | 74.41 ms | 24 - 38 MB | FP16 | NPU | Use Export Script |
55
- | FFNet-78S | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 50.659 ms | 2 - 25 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
56
- | FFNet-78S | SA8650 (Proxy) | SA8650P Proxy | QNN | 46.988 ms | 24 - 27 MB | FP16 | NPU | Use Export Script |
57
- | FFNet-78S | SA8775P ADP | SA8775P | TFLITE | 86.094 ms | 2 - 54 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
58
- | FFNet-78S | SA8775P ADP | SA8775P | QNN | 78.862 ms | 24 - 34 MB | FP16 | NPU | Use Export Script |
59
- | FFNet-78S | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 68.751 ms | 2 - 49 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
60
- | FFNet-78S | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 80.15 ms | 24 - 65 MB | FP16 | NPU | Use Export Script |
61
- | FFNet-78S | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 47.54 ms | 24 - 24 MB | FP16 | NPU | Use Export Script |
62
- | FFNet-78S | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 44.3 ms | 31 - 31 MB | FP16 | NPU | [FFNet-78S.onnx](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.onnx) |
63
 
64
 
65
 
66
 
67
  ## Installation
68
 
69
- This model can be installed as a Python package via pip.
70
 
 
71
  ```bash
72
- pip install "qai-hub-models[ffnet_78s]"
73
  ```
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
@@ -124,8 +123,8 @@ Profiling Results
124
  FFNet-78S
125
  Device : Samsung Galaxy S23 (13)
126
  Runtime : TFLITE
127
- Estimated inference time (ms) : 50.7
128
- Estimated peak memory usage (MB): [2, 34]
129
  Total # Ops : 151
130
  Compute Unit(s) : NPU (151 ops)
131
  ```
@@ -152,7 +151,7 @@ from qai_hub_models.models.ffnet_78s import Model
152
  torch_model = Model.from_pretrained()
153
 
154
  # Device
155
- device = hub.Device("Samsung Galaxy S23")
156
 
157
  # Trace model
158
  input_shape = torch_model.get_input_spec()
@@ -244,7 +243,8 @@ Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)
244
 
245
 
246
  ## License
247
- * The license for the original implementation of FFNet-78S can be found [here](https://github.com/Qualcomm-AI-research/FFNet/blob/master/LICENSE).
 
248
  * 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)
249
 
250
 
 
35
 
36
  | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
38
+ | FFNet-78S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 50.784 ms | 2 - 32 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
39
+ | FFNet-78S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 61.079 ms | 24 - 43 MB | FP16 | NPU | [FFNet-78S.so](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.so) |
40
+ | FFNet-78S | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 39.841 ms | 5 - 166 MB | FP16 | NPU | [FFNet-78S.onnx](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.onnx) |
41
+ | FFNet-78S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 34.461 ms | 0 - 54 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
42
+ | FFNet-78S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 41.668 ms | 24 - 67 MB | FP16 | NPU | [FFNet-78S.so](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.so) |
43
+ | FFNet-78S | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 32.25 ms | 30 - 91 MB | FP16 | NPU | [FFNet-78S.onnx](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.onnx) |
44
+ | FFNet-78S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 27.435 ms | 1 - 55 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
45
+ | FFNet-78S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 44.458 ms | 24 - 71 MB | FP16 | NPU | Use Export Script |
46
+ | FFNet-78S | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 29.263 ms | 24 - 77 MB | FP16 | NPU | [FFNet-78S.onnx](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.onnx) |
47
+ | FFNet-78S | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 50.816 ms | 2 - 27 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
48
+ | FFNet-78S | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 47.059 ms | 24 - 27 MB | FP16 | NPU | Use Export Script |
49
+ | FFNet-78S | SA7255P ADP | SA7255P | TFLITE | 1255.948 ms | 0 - 51 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
50
+ | FFNet-78S | SA7255P ADP | SA7255P | QNN | 1243.805 ms | 24 - 32 MB | FP16 | NPU | Use Export Script |
51
+ | FFNet-78S | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 50.912 ms | 2 - 34 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
52
+ | FFNet-78S | SA8255 (Proxy) | SA8255P Proxy | QNN | 47.189 ms | 24 - 27 MB | FP16 | NPU | Use Export Script |
53
+ | FFNet-78S | SA8295P ADP | SA8295P | TFLITE | 78.897 ms | 2 - 49 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
54
+ | FFNet-78S | SA8295P ADP | SA8295P | QNN | 74.199 ms | 24 - 38 MB | FP16 | NPU | Use Export Script |
55
+ | FFNet-78S | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 50.693 ms | 2 - 27 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
56
+ | FFNet-78S | SA8650 (Proxy) | SA8650P Proxy | QNN | 47.286 ms | 24 - 27 MB | FP16 | NPU | Use Export Script |
57
+ | FFNet-78S | SA8775P ADP | SA8775P | TFLITE | 86.123 ms | 3 - 54 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
58
+ | FFNet-78S | SA8775P ADP | SA8775P | QNN | 78.836 ms | 24 - 34 MB | FP16 | NPU | Use Export Script |
59
+ | FFNet-78S | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 68.773 ms | 2 - 48 MB | FP16 | NPU | [FFNet-78S.tflite](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.tflite) |
60
+ | FFNet-78S | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 80.603 ms | 22 - 65 MB | FP16 | NPU | Use Export Script |
61
+ | FFNet-78S | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 47.452 ms | 24 - 24 MB | FP16 | NPU | Use Export Script |
62
+ | FFNet-78S | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 44.519 ms | 31 - 31 MB | FP16 | NPU | [FFNet-78S.onnx](https://huggingface.co/qualcomm/FFNet-78S/blob/main/FFNet-78S.onnx) |
63
 
64
 
65
 
66
 
67
  ## Installation
68
 
 
69
 
70
+ Install the package via pip:
71
  ```bash
72
+ pip install "qai-hub-models[ffnet-78s]"
73
  ```
74
 
75
 
 
76
  ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
77
 
78
  Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
 
123
  FFNet-78S
124
  Device : Samsung Galaxy S23 (13)
125
  Runtime : TFLITE
126
+ Estimated inference time (ms) : 50.8
127
+ Estimated peak memory usage (MB): [2, 32]
128
  Total # Ops : 151
129
  Compute Unit(s) : NPU (151 ops)
130
  ```
 
151
  torch_model = Model.from_pretrained()
152
 
153
  # Device
154
+ device = hub.Device("Samsung Galaxy S24")
155
 
156
  # Trace model
157
  input_shape = torch_model.get_input_spec()
 
243
 
244
 
245
  ## License
246
+ * The license for the original implementation of FFNet-78S can be found
247
+ [here](https://github.com/Qualcomm-AI-research/FFNet/blob/master/LICENSE).
248
  * 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)
249
 
250