qaihm-bot commited on
Commit
820743b
·
verified ·
1 Parent(s): 90304e7

See https://github.com/quic/ai-hub-models/releases/v0.39.0 for changelog.

README.md CHANGED
@@ -37,33 +37,38 @@ More details on model performance across various devices, can be found
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
  | Sequencer2D | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 116.656 ms | 0 - 542 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
40
- | Sequencer2D | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 59.22 ms | 1 - 805 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
41
  | Sequencer2D | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 57.773 ms | 0 - 411 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
42
- | Sequencer2D | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 67.919 ms | 1 - 432 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
43
  | Sequencer2D | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 60.096 ms | 0 - 70 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
44
- | Sequencer2D | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 22.679 ms | 0 - 107 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
45
  | Sequencer2D | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 46.962 ms | 0 - 238 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.onnx.zip) |
46
  | Sequencer2D | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 62.087 ms | 0 - 541 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
 
47
  | Sequencer2D | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 43.924 ms | 0 - 543 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
48
- | Sequencer2D | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 15.732 ms | 1 - 1656 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
49
  | Sequencer2D | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 33.638 ms | 1 - 872 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.onnx.zip) |
50
  | Sequencer2D | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 37.148 ms | 0 - 537 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
51
- | Sequencer2D | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 14.601 ms | 0 - 803 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
52
  | Sequencer2D | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 23.506 ms | 1 - 508 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.onnx.zip) |
53
- | Sequencer2D | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 23.775 ms | 830 - 830 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
 
 
54
  | Sequencer2D | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 48.364 ms | 65 - 65 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.onnx.zip) |
55
- | Sequencer2D | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 82.339 ms | 0 - 486 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
56
- | Sequencer2D | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 39.607 ms | 0 - 381 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
57
- | Sequencer2D | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 44.655 ms | 0 - 67 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
58
- | Sequencer2D | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 84.222 ms | 121 - 257 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
59
- | Sequencer2D | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 45.772 ms | 0 - 485 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
60
- | Sequencer2D | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 262.468 ms | 20 - 41 MB | CPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
61
- | Sequencer2D | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 246.974 ms | 17 - 48 MB | CPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
62
- | Sequencer2D | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 32.408 ms | 0 - 490 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
63
- | Sequencer2D | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 65.763 ms | 137 - 2462 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
64
- | Sequencer2D | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 26.784 ms | 0 - 480 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
65
- | Sequencer2D | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 58.447 ms | 160 - 712 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
66
- | Sequencer2D | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 80.04 ms | 232 - 232 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
 
 
 
67
 
68
 
69
 
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
  | Sequencer2D | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 116.656 ms | 0 - 542 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
 
40
  | Sequencer2D | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 57.773 ms | 0 - 411 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
41
+ | Sequencer2D | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 68.072 ms | 1 - 431 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
42
  | Sequencer2D | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 60.096 ms | 0 - 70 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
43
+ | Sequencer2D | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 22.841 ms | 0 - 105 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
44
  | Sequencer2D | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 46.962 ms | 0 - 238 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.onnx.zip) |
45
  | Sequencer2D | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 62.087 ms | 0 - 541 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
46
+ | Sequencer2D | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 26.485 ms | 36 - 1231 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
47
  | Sequencer2D | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 43.924 ms | 0 - 543 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
48
+ | Sequencer2D | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 15.684 ms | 0 - 1655 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
49
  | Sequencer2D | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 33.638 ms | 1 - 872 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.onnx.zip) |
50
  | Sequencer2D | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 37.148 ms | 0 - 537 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
51
+ | Sequencer2D | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 14.475 ms | 0 - 804 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
52
  | Sequencer2D | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 23.506 ms | 1 - 508 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.onnx.zip) |
53
+ | Sequencer2D | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 35.766 ms | 0 - 499 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.tflite) |
54
+ | Sequencer2D | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 18.535 ms | 1 - 937 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.onnx.zip) |
55
+ | Sequencer2D | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 23.723 ms | 832 - 832 MB | NPU | [Sequencer2D.dlc](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.dlc) |
56
  | Sequencer2D | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 48.364 ms | 65 - 65 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D.onnx.zip) |
57
+ | Sequencer2D | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 82.278 ms | 0 - 493 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
58
+ | Sequencer2D | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 39.8 ms | 0 - 381 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
59
+ | Sequencer2D | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 44.427 ms | 0 - 73 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
60
+ | Sequencer2D | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 89.188 ms | 121 - 255 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
61
+ | Sequencer2D | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 45.857 ms | 0 - 490 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
62
+ | Sequencer2D | w8a8 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 279.052 ms | 19 - 40 MB | CPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
63
+ | Sequencer2D | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 242.143 ms | 13 - 45 MB | CPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
64
+ | Sequencer2D | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 32.384 ms | 0 - 491 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
65
+ | Sequencer2D | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 65.559 ms | 163 - 766 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
66
+ | Sequencer2D | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 26.595 ms | 0 - 475 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
67
+ | Sequencer2D | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 46.191 ms | 161 - 715 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
68
+ | Sequencer2D | w8a8 | Snapdragon 7 Gen 5 QRD | Snapdragon® 7 Gen 5 Mobile | ONNX | 247.8 ms | 19 - 38 MB | CPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
69
+ | Sequencer2D | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 25.13 ms | 0 - 478 MB | NPU | [Sequencer2D.tflite](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.tflite) |
70
+ | Sequencer2D | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 43.23 ms | 147 - 1039 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
71
+ | Sequencer2D | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 79.085 ms | 228 - 228 MB | NPU | [Sequencer2D.onnx.zip](https://huggingface.co/qualcomm/Sequencer2D/blob/main/Sequencer2D_w8a8.onnx.zip) |
72
 
73
 
74
 
Sequencer2D_float.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a9db2adcd035c34de7afc1743501f6c585e402a0b5c8519461d0dee33b4440de
3
- size 123433188
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28483a5432867f4f83889eb555cb566b6a460e8b5bef7ed630f02682939c0001
3
+ size 123433740
Sequencer2D_w8a8.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e3c95075b3b6a958bedf05f20e7d4bd984dcb15c39236f6e24261012016ebb45
3
- size 72139773
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ae8276e09a986b11a113cf1eecc3fd442e0c209ee775a51b49c4d8686088bff4
3
+ size 72139775
Sequencer2D_w8a8.tflite CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:23992fc4fd4ffc02e8314f7809df8bbb0b363520c5ba2f5631494a287a7374b0
3
  size 65551016
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3e2e452e76fe4d7b0f298930a08309d38cc8837918c00ab9b6f49d0bffae561
3
  size 65551016