Hugging Face's logo

qualcomm
/
DeepLabV3-ResNet50

Image Segmentation
PyTorch
LiteRT
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
152903c
·
verified ·
1 Parent(s): b7f995e

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +21 -21
README.md CHANGED
@@ -1,6 +1,6 @@
1
  ---
2
  library_name: pytorch
3
- license: bsd-3-clause
4
  tags:
5
  - android
6
  pipeline_tag: image-segmentation
@@ -25,7 +25,7 @@ More details on model performance across various devices, can be found
25
 
26
  ### Model Details
27
 
28
- - **Model Type:** Semantic segmentation
29
  - **Model Stats:**
30
  - Model checkpoint: COCO_WITH_VOC_LABELS_V1
31
  - Input resolution: 513x513
@@ -33,20 +33,20 @@ More details on model performance across various devices, can be found
33
  - Model size: 151 MB
34
  - Number of output classes: 21
35
 
36
- | Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
38
- | DeepLabV3-ResNet50 | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 294.587 ms | 0 - 159 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
39
- | DeepLabV3-ResNet50 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 229.213 ms | 21 - 50 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
40
- | DeepLabV3-ResNet50 | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 220.99 ms | 22 - 41 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
41
- | DeepLabV3-ResNet50 | SA7255P ADP | SA7255P | TFLITE | 2156.441 ms | 21 - 39 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
42
- | DeepLabV3-ResNet50 | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 292.574 ms | 0 - 189 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
43
- | DeepLabV3-ResNet50 | SA8295P ADP | SA8295P | TFLITE | 279.912 ms | 23 - 44 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
44
- | DeepLabV3-ResNet50 | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 293.467 ms | 2 - 177 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
45
- | DeepLabV3-ResNet50 | SA8775P ADP | SA8775P | TFLITE | 592.126 ms | 22 - 42 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
46
- | DeepLabV3-ResNet50 | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 2156.441 ms | 21 - 39 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
47
- | DeepLabV3-ResNet50 | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 291.59 ms | 0 - 234 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
48
- | DeepLabV3-ResNet50 | QCS9075 (Proxy) | QCS9075 Proxy | TFLITE | 592.126 ms | 22 - 42 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
49
- | DeepLabV3-ResNet50 | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 1424.905 ms | 16 - 48 MB | FP16 | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
50
 
51
 
52
 
@@ -108,12 +108,12 @@ python -m qai_hub_models.models.deeplabv3_resnet50.export
108
  Profiling Results
109
  ------------------------------------------------------------
110
  DeepLabV3-ResNet50
111
- Device : Samsung Galaxy S23 (13)
112
- Runtime : TFLITE
113
- Estimated inference time (ms) : 294.6
114
- Estimated peak memory usage (MB): [0, 159]
115
- Total # Ops : 100
116
- Compute Unit(s) : GPU (98 ops) CPU (2 ops)
117
  ```
118
 
119
 
 
1
  ---
2
  library_name: pytorch
3
+ license: other
4
  tags:
5
  - android
6
  pipeline_tag: image-segmentation
 
25
 
26
  ### Model Details
27
 
28
+ - **Model Type:** Model_use_case.semantic_segmentation
29
  - **Model Stats:**
30
  - Model checkpoint: COCO_WITH_VOC_LABELS_V1
31
  - Input resolution: 513x513
 
33
  - Model size: 151 MB
34
  - Number of output classes: 21
35
 
36
+ | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
38
+ | DeepLabV3-ResNet50 | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 1554.729 ms | 20 - 36 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
39
+ | DeepLabV3-ResNet50 | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 434.261 ms | 23 - 59 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
40
+ | DeepLabV3-ResNet50 | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 295.627 ms | 2 - 197 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
41
+ | DeepLabV3-ResNet50 | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 593.065 ms | 22 - 39 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
42
+ | DeepLabV3-ResNet50 | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 1554.729 ms | 20 - 36 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
43
+ | DeepLabV3-ResNet50 | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 294.807 ms | 2 - 200 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
44
+ | DeepLabV3-ResNet50 | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 265.22 ms | 23 - 44 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
45
+ | DeepLabV3-ResNet50 | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 291.339 ms | 2 - 206 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
46
+ | DeepLabV3-ResNet50 | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 593.065 ms | 22 - 39 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
47
+ | DeepLabV3-ResNet50 | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 291.57 ms | 2 - 236 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
48
+ | DeepLabV3-ResNet50 | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 298.315 ms | 20 - 54 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
49
+ | DeepLabV3-ResNet50 | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 188.721 ms | 21 - 41 MB | GPU | [DeepLabV3-ResNet50.tflite](https://huggingface.co/qualcomm/DeepLabV3-ResNet50/blob/main/DeepLabV3-ResNet50.tflite) |
50
 
51
 
52
 
 
108
  Profiling Results
109
  ------------------------------------------------------------
110
  DeepLabV3-ResNet50
111
+ Device : cs_8275 (ANDROID 14)
112
+ Runtime : TFLITE
113
+ Estimated inference time (ms) : 1554.7
114
+ Estimated peak memory usage (MB): [20, 36]
115
+ Total # Ops : 100
116
+ Compute Unit(s) : npu (0 ops) gpu (98 ops) cpu (2 ops)
117
  ```
118
 
119