qaihm-bot commited on
Commit
4ad85d6
·
verified ·
1 Parent(s): 6abb827

See https://github.com/quic/ai-hub-models/releases/v0.31.0 for changelog.

Files changed (3) hide show
  1. .gitattributes +1 -0
  2. DLA-102-X.dlc +3 -0
  3. README.md +19 -20
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ DLA-102-X.dlc filter=lfs diff=lfs merge=lfs -text
DLA-102-X.dlc ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:512e33303510005e4710c9d289ed16d36e19378dfa5e0abdf88cc2f470ce1156
3
+ size 105348251
README.md CHANGED
@@ -35,25 +35,24 @@ 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
- | DLA-102-X | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 14.59 ms | 0 - 115 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
39
- | DLA-102-X | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN | 104.26 ms | 1 - 10 MB | NPU | Use Export Script |
40
- | DLA-102-X | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 3.827 ms | 0 - 105 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
41
- | DLA-102-X | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN | 4.346 ms | 1 - 39 MB | NPU | Use Export Script |
42
- | DLA-102-X | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 2.922 ms | 0 - 99 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
43
- | DLA-102-X | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN | 2.859 ms | 1 - 4 MB | NPU | Use Export Script |
44
- | DLA-102-X | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 4.505 ms | 0 - 115 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
45
- | DLA-102-X | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN | 4.401 ms | 1 - 10 MB | NPU | Use Export Script |
46
- | DLA-102-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 2.929 ms | 0 - 28 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
47
- | DLA-102-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN | 2.812 ms | 0 - 20 MB | NPU | Use Export Script |
48
- | DLA-102-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 2.907 ms | 0 - 173 MB | NPU | [DLA-102-X.onnx](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.onnx) |
49
- | DLA-102-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 2.034 ms | 0 - 117 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
50
- | DLA-102-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN | 1.962 ms | 1 - 55 MB | NPU | Use Export Script |
51
- | DLA-102-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.051 ms | 0 - 60 MB | NPU | [DLA-102-X.onnx](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.onnx) |
52
- | DLA-102-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 1.662 ms | 0 - 116 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
53
- | DLA-102-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN | 1.9 ms | 1 - 51 MB | NPU | Use Export Script |
54
- | DLA-102-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.792 ms | 1 - 52 MB | NPU | [DLA-102-X.onnx](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.onnx) |
55
- | DLA-102-X | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 3.095 ms | 1 - 1 MB | NPU | Use Export Script |
56
- | DLA-102-X | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.867 ms | 54 - 54 MB | NPU | [DLA-102-X.onnx](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.onnx) |
57
 
58
 
59
 
@@ -118,7 +117,7 @@ DLA-102-X
118
  Device : cs_8275 (ANDROID 14)
119
  Runtime : TFLITE
120
  Estimated inference time (ms) : 14.6
121
- Estimated peak memory usage (MB): [0, 115]
122
  Total # Ops : 175
123
  Compute Unit(s) : npu (175 ops) gpu (0 ops) cpu (0 ops)
124
  ```
 
35
 
36
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
37
  |---|---|---|---|---|---|---|---|---|
38
+ | DLA-102-X | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 14.578 ms | 0 - 114 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
39
+ | DLA-102-X | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 14.339 ms | 1 - 48 MB | NPU | [DLA-102-X.dlc](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.dlc) |
40
+ | DLA-102-X | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 3.803 ms | 0 - 107 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
41
+ | DLA-102-X | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 4.412 ms | 1 - 40 MB | NPU | [DLA-102-X.dlc](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.dlc) |
42
+ | DLA-102-X | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 2.829 ms | 0 - 130 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
43
+ | DLA-102-X | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 2.768 ms | 2 - 22 MB | NPU | [DLA-102-X.dlc](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.dlc) |
44
+ | DLA-102-X | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 4.539 ms | 0 - 114 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
45
+ | DLA-102-X | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 4.34 ms | 1 - 48 MB | NPU | [DLA-102-X.dlc](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.dlc) |
46
+ | DLA-102-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 2.865 ms | 0 - 60 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
47
+ | DLA-102-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 2.787 ms | 0 - 20 MB | NPU | [DLA-102-X.dlc](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.dlc) |
48
+ | DLA-102-X | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 2.909 ms | 0 - 174 MB | NPU | [DLA-102-X.onnx](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.onnx) |
49
+ | DLA-102-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 2.034 ms | 0 - 126 MB | NPU | [DLA-102-X.tflite](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.tflite) |
50
+ | DLA-102-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.97 ms | 0 - 56 MB | NPU | [DLA-102-X.dlc](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.dlc) |
51
+ | DLA-102-X | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.121 ms | 0 - 64 MB | NPU | [DLA-102-X.onnx](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.onnx) |
52
+ | DLA-102-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 1.898 ms | 1 - 51 MB | NPU | [DLA-102-X.dlc](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.dlc) |
53
+ | DLA-102-X | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 2.075 ms | 1 - 52 MB | NPU | [DLA-102-X.onnx](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.onnx) |
54
+ | DLA-102-X | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 3.458 ms | 156 - 156 MB | NPU | [DLA-102-X.dlc](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.dlc) |
55
+ | DLA-102-X | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.895 ms | 54 - 54 MB | NPU | [DLA-102-X.onnx](https://huggingface.co/qualcomm/DLA-102-X/blob/main/DLA-102-X.onnx) |
 
56
 
57
 
58
 
 
117
  Device : cs_8275 (ANDROID 14)
118
  Runtime : TFLITE
119
  Estimated inference time (ms) : 14.6
120
+ Estimated peak memory usage (MB): [0, 114]
121
  Total # Ops : 175
122
  Compute Unit(s) : npu (175 ops) gpu (0 ops) cpu (0 ops)
123
  ```