Hugging Face's logo

qualcomm
/
Video-MAE

Video Classification
PyTorch
LiteRT
backbone
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
feb3a6c
·
verified ·
1 Parent(s): 170a6c7

v0.38.0

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.38.0 for changelog.

Files changed (7) hide show
  1. README.md +27 -30
  2. Video-MAE_float.dlc +2 -2
  3. Video-MAE_float.onnx.zip +2 -2
  4. precompiled/qualcomm-snapdragon-x-elite/Video-MAE_float.bin +0 -3
  5. precompiled/qualcomm-snapdragon-x-elite/Video-MAE_float.onnx.zip +0 -3
  6. precompiled/qualcomm-snapdragon-x-elite/tool-versions.yaml +0 -3
  7. tool-versions.yaml +2 -2
README.md CHANGED
@@ -36,35 +36,32 @@ More details on model performance across various devices, can be found
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
- | Video-MAE | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 442.754 ms | 0 - 522 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
40
- | Video-MAE | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 1205.047 ms | 3 - 552 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
41
- | Video-MAE | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 219.79 ms | 3 - 486 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
42
- | Video-MAE | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1403.109 ms | 9 - 427 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
43
- | Video-MAE | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 149.154 ms | 0 - 37 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
44
- | Video-MAE | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 469.676 ms | 9 - 45 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
45
- | Video-MAE | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 585.788 ms | 0 - 249 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
46
- | Video-MAE | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 171.607 ms | 0 - 522 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
47
- | Video-MAE | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 512.953 ms | 2 - 455 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
48
- | Video-MAE | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 442.754 ms | 0 - 522 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
49
- | Video-MAE | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 1205.047 ms | 3 - 552 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
50
- | Video-MAE | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 149.07 ms | 0 - 36 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
51
- | Video-MAE | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 469.042 ms | 9 - 48 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
52
- | Video-MAE | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 239.774 ms | 0 - 480 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
53
- | Video-MAE | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 740.755 ms | 6 - 423 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
54
- | Video-MAE | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 148.626 ms | 0 - 36 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
55
- | Video-MAE | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 469.418 ms | 9 - 49 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
56
- | Video-MAE | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 171.607 ms | 0 - 522 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
57
- | Video-MAE | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 512.953 ms | 2 - 455 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
58
- | Video-MAE | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 148.891 ms | 0 - 44 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
59
- | Video-MAE | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 469.928 ms | 9 - 47 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
60
- | Video-MAE | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 586.169 ms | 9 - 48 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
61
- | Video-MAE | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 109.661 ms | 46 - 568 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
62
- | Video-MAE | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 372.772 ms | 9 - 565 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
63
- | Video-MAE | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 386.596 ms | 9 - 489 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
64
- | Video-MAE | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 337.054 ms | 1 - 489 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
65
- | Video-MAE | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 574.665 ms | 9 - 530 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
66
- | Video-MAE | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 483.963 ms | 636 - 636 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
67
- | Video-MAE | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 589.06 ms | 188 - 188 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
68
 
69
 
70
 
@@ -146,7 +143,7 @@ from qai_hub_models.models.video_mae import Model
146
  torch_model = Model.from_pretrained()
147
 
148
  # Device
149
- device = hub.Device("Samsung Galaxy S24")
150
 
151
  # Trace model
152
  input_shape = torch_model.get_input_spec()
 
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
+ | Video-MAE | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 442.791 ms | 0 - 516 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
40
+ | Video-MAE | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 1207.297 ms | 5 - 554 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
41
+ | Video-MAE | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 221.592 ms | 0 - 486 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
42
+ | Video-MAE | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1403.99 ms | 9 - 433 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
43
+ | Video-MAE | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 148.909 ms | 0 - 41 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
44
+ | Video-MAE | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 469.685 ms | 9 - 43 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
45
+ | Video-MAE | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 581.607 ms | 9 - 51 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
46
+ | Video-MAE | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 171.58 ms | 0 - 522 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
47
+ | Video-MAE | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 513.909 ms | 4 - 458 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
48
+ | Video-MAE | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 442.791 ms | 0 - 516 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
49
+ | Video-MAE | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 1207.297 ms | 5 - 554 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
50
+ | Video-MAE | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 149.361 ms | 0 - 43 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
51
+ | Video-MAE | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 469.547 ms | 9 - 52 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
52
+ | Video-MAE | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 742.468 ms | 6 - 425 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
53
+ | Video-MAE | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 150.074 ms | 0 - 45 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
54
+ | Video-MAE | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 469.235 ms | 9 - 68 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
55
+ | Video-MAE | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 171.58 ms | 0 - 522 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
56
+ | Video-MAE | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 513.909 ms | 4 - 458 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
57
+ | Video-MAE | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 109.476 ms | 0 - 521 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
58
+ | Video-MAE | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 373.376 ms | 9 - 569 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
59
+ | Video-MAE | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 421.448 ms | 0 - 576 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
60
+ | Video-MAE | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 81.878 ms | 0 - 488 MB | NPU | [Video-MAE.tflite](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.tflite) |
61
+ | Video-MAE | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 279.105 ms | 9 - 500 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
62
+ | Video-MAE | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 441.895 ms | 0 - 577 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
63
+ | Video-MAE | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 484.049 ms | 547 - 547 MB | NPU | [Video-MAE.dlc](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.dlc) |
64
+ | Video-MAE | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 580.976 ms | 187 - 187 MB | NPU | [Video-MAE.onnx.zip](https://huggingface.co/qualcomm/Video-MAE/blob/main/Video-MAE.onnx.zip) |
 
 
 
65
 
66
 
67
 
 
143
  torch_model = Model.from_pretrained()
144
 
145
  # Device
146
+ device = hub.Device("Samsung Galaxy S25")
147
 
148
  # Trace model
149
  input_shape = torch_model.get_input_spec()
Video-MAE_float.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:08128dba2cb2a038680b847b8d9590ea83bb07ce070c7cecf0d3f915c91561c2
3
- size 351321508
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d30e86cea74dd7535a08df937a6df77b08943e120d5b608810deb50697c82c45
3
+ size 351321588
Video-MAE_float.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:887ee01fe26c7cf8f2dbfbdd3b098bd88c23439dbabe239366c7ee4b682ee936
3
- size 221433191
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d78d33baed792522b5e8550223ab7e6a036ffada2744ac95255da4576ba0c50e
3
+ size 221433387
precompiled/qualcomm-snapdragon-x-elite/Video-MAE_float.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:28bc782a8a51d49c9fef6d0073b1335632bf9c471315be079a2e1c5effee5652
3
- size 205021184
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/Video-MAE_float.onnx.zip DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:081d571ff7e901e4d3544d143ab06e555fd90e84f87e4dce48b71f21f6e5e846
3
- size 165422107
 
 
 
 
precompiled/qualcomm-snapdragon-x-elite/tool-versions.yaml DELETED
@@ -1,3 +0,0 @@
1
- tool_versions:
2
- precompiled_qnn_onnx:
3
- qairt: 2.36.4.250725200057_123280
 
 
 
 
tool-versions.yaml CHANGED
@@ -1,4 +1,4 @@
1
  tool_versions:
2
  onnx:
3
- qairt: 2.36.4.250725200057_123280
4
- onnx_runtime: 1.22.0
 
1
  tool_versions:
2
  onnx:
3
+ qairt: 2.37.1.250807093845_124904
4
+ onnx_runtime: 1.22.2