Hugging Face's logo

qualcomm
/
FastSam-X

Image Segmentation
PyTorch
LiteRT
android
Model card Files Community
qaihm-bot commited on
Commit
c1854e4
·
verified ·
1 Parent(s): b4adf16

v0.30.5

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.30.5 for changelog.

Files changed (1) hide show
  1. README.md +34 -36
README.md CHANGED
@@ -35,35 +35,33 @@ More details on model performance across various devices, can be found
35
 
36
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
38
- | FastSam-X | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 280.693 ms | 4 - 98 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
39
- | FastSam-X | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN | 277.43 ms | 0 - 9 MB | NPU | Use Export Script |
40
- | FastSam-X | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 86.8 ms | 4 - 193 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
41
- | FastSam-X | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN | 100.032 ms | 5 - 76 MB | NPU | Use Export Script |
42
- | FastSam-X | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 45.231 ms | 4 - 57 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
43
- | FastSam-X | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN | 42.906 ms | 5 - 7 MB | NPU | Use Export Script |
44
- | FastSam-X | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 70.591 ms | 4 - 98 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
45
- | FastSam-X | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN | 68.099 ms | 1 - 11 MB | NPU | Use Export Script |
46
- | FastSam-X | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 280.693 ms | 4 - 98 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
47
- | FastSam-X | float | SA7255P ADP | Qualcomm® SA7255P | QNN | 277.43 ms | 0 - 9 MB | NPU | Use Export Script |
48
- | FastSam-X | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 47.383 ms | 3 - 58 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
49
- | FastSam-X | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN | 43.695 ms | 5 - 7 MB | NPU | Use Export Script |
50
- | FastSam-X | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 89.652 ms | 4 - 95 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
51
- | FastSam-X | float | SA8295P ADP | Qualcomm® SA8295P | QNN | 80.068 ms | 0 - 18 MB | NPU | Use Export Script |
52
- | FastSam-X | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 45.586 ms | 4 - 58 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
53
- | FastSam-X | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN | 42.452 ms | 5 - 7 MB | NPU | Use Export Script |
54
- | FastSam-X | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 70.591 ms | 4 - 98 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
55
- | FastSam-X | float | SA8775P ADP | Qualcomm® SA8775P | QNN | 68.099 ms | 1 - 11 MB | NPU | Use Export Script |
56
- | FastSam-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 45.519 ms | 4 - 58 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
57
- | FastSam-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN | 42.512 ms | 5 - 32 MB | NPU | Use Export Script |
58
- | FastSam-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 46.427 ms | 12 - 326 MB | NPU | [FastSam-X.onnx](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.onnx) |
59
- | FastSam-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 34.898 ms | 3 - 190 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
60
- | FastSam-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN | 32.512 ms | 5 - 64 MB | NPU | Use Export Script |
61
- | FastSam-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 33.508 ms | 17 - 82 MB | NPU | [FastSam-X.onnx](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.onnx) |
62
- | FastSam-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 31.028 ms | 4 - 100 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
63
- | FastSam-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN | 25.358 ms | 5 - 59 MB | NPU | Use Export Script |
64
- | FastSam-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 30.147 ms | 15 - 72 MB | NPU | [FastSam-X.onnx](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.onnx) |
65
- | FastSam-X | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 43.446 ms | 5 - 5 MB | NPU | Use Export Script |
66
- | FastSam-X | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 46.969 ms | 139 - 139 MB | NPU | [FastSam-X.onnx](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.onnx) |
67
 
68
 
69
 
@@ -126,11 +124,11 @@ Profiling Results
126
  ------------------------------------------------------------
127
  FastSam-X
128
  Device : cs_8275 (ANDROID 14)
129
- Runtime : TFLITE
130
- Estimated inference time (ms) : 280.7
131
- Estimated peak memory usage (MB): [4, 98]
132
- Total # Ops : 419
133
- Compute Unit(s) : npu (419 ops) gpu (0 ops) cpu (0 ops)
134
  ```
135
 
136
 
@@ -217,13 +215,13 @@ AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
217
  You can also run the demo on-device.
218
 
219
  ```bash
220
- python -m qai_hub_models.models.fastsam_x.demo --on-device
221
  ```
222
 
223
  **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
224
  environment, please add the following to your cell (instead of the above).
225
  ```
226
- %run -m qai_hub_models.models.fastsam_x.demo -- --on-device
227
  ```
228
 
229
 
 
35
 
36
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
38
+ | FastSam-X | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN | 2094.279 ms | 5 - 14 MB | NPU | Use Export Script |
39
+ | FastSam-X | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 92.958 ms | 4 - 199 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
40
+ | FastSam-X | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN | 99.394 ms | 5 - 88 MB | NPU | Use Export Script |
41
+ | FastSam-X | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 43.602 ms | 4 - 68 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
42
+ | FastSam-X | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN | 42.594 ms | 5 - 8 MB | NPU | Use Export Script |
43
+ | FastSam-X | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 70.714 ms | 4 - 118 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
44
+ | FastSam-X | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN | 68.128 ms | 1 - 11 MB | NPU | Use Export Script |
45
+ | FastSam-X | float | SA7255P ADP | Qualcomm® SA7255P | QNN | 2094.279 ms | 5 - 14 MB | NPU | Use Export Script |
46
+ | FastSam-X | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 45.407 ms | 1 - 60 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
47
+ | FastSam-X | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN | 43.169 ms | 5 - 7 MB | NPU | Use Export Script |
48
+ | FastSam-X | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 89.577 ms | 3 - 104 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
49
+ | FastSam-X | float | SA8295P ADP | Qualcomm® SA8295P | QNN | 79.849 ms | 0 - 16 MB | NPU | Use Export Script |
50
+ | FastSam-X | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 46.386 ms | 4 - 58 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
51
+ | FastSam-X | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN | 42.826 ms | 5 - 7 MB | NPU | Use Export Script |
52
+ | FastSam-X | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 70.714 ms | 4 - 118 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
53
+ | FastSam-X | float | SA8775P ADP | Qualcomm® SA8775P | QNN | 68.128 ms | 1 - 11 MB | NPU | Use Export Script |
54
+ | FastSam-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 43.786 ms | 4 - 60 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
55
+ | FastSam-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN | 43.288 ms | 5 - 31 MB | NPU | Use Export Script |
56
+ | FastSam-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 48.573 ms | 12 - 322 MB | NPU | [FastSam-X.onnx](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.onnx) |
57
+ | FastSam-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 34.806 ms | 3 - 213 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
58
+ | FastSam-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN | 33.533 ms | 5 - 89 MB | NPU | Use Export Script |
59
+ | FastSam-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 34.229 ms | 13 - 99 MB | NPU | [FastSam-X.onnx](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.onnx) |
60
+ | FastSam-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 30.673 ms | 3 - 120 MB | NPU | [FastSam-X.tflite](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.tflite) |
61
+ | FastSam-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN | 28.818 ms | 5 - 83 MB | NPU | Use Export Script |
62
+ | FastSam-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 26.71 ms | 17 - 90 MB | NPU | [FastSam-X.onnx](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.onnx) |
63
+ | FastSam-X | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 43.717 ms | 5 - 5 MB | NPU | Use Export Script |
64
+ | FastSam-X | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 47.241 ms | 139 - 139 MB | NPU | [FastSam-X.onnx](https://huggingface.co/qualcomm/FastSam-X/blob/main/FastSam-X.onnx) |
 
 
65
 
66
 
67
 
 
124
  ------------------------------------------------------------
125
  FastSam-X
126
  Device : cs_8275 (ANDROID 14)
127
+ Runtime : QNN
128
+ Estimated inference time (ms) : 2094.3
129
+ Estimated peak memory usage (MB): [5, 14]
130
+ Total # Ops : 418
131
+ Compute Unit(s) : npu (418 ops) gpu (0 ops) cpu (0 ops)
132
  ```
133
 
134
 
 
215
  You can also run the demo on-device.
216
 
217
  ```bash
218
+ python -m qai_hub_models.models.fastsam_x.demo --eval-mode on-device
219
  ```
220
 
221
  **NOTE**: If you want running in a Jupyter Notebook or Google Colab like
222
  environment, please add the following to your cell (instead of the above).
223
  ```
224
+ %run -m qai_hub_models.models.fastsam_x.demo -- --eval-mode on-device
225
  ```
226
 
227