Hugging Face's logo

qualcomm
/
ConvNext-Base

Image Classification
PyTorch
LiteRT
bu_auto
android
Model card Files
xet
Community
qaihm-bot commited on
Commit
30def4b
·
verified ·
1 Parent(s): 4dc3363

v0.42.0

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.42.0 for changelog.

Files changed (5) hide show
  1. ConvNext-Base_float.dlc +2 -2
  2. ConvNext-Base_float.onnx.zip +1 -1
  3. ConvNext-Base_w8a16.dlc +2 -2
  4. ConvNext-Base_w8a16.onnx.zip +2 -2
  5. README.md +40 -38
ConvNext-Base_float.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9b85c3767adacb5bf3fa1a956dd3e11b84e1df68820875e1038e0f5c6fa49096
3
- size 354737300
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bfa4f5898620df4ca1e0c466a8c698cd2e9b712ca87c0794e2b4a4e5879a1947
3
+ size 354737404
ConvNext-Base_float.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:63ea39bac75b29dc3768a076c7037aabe08b043244deb697f24344e951a10e52
3
  size 329653286
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:627f693ff39a28a6d5fdf957a334b704fc7d61eac3dd08e61258a1dd2a2a9bc9
3
  size 329653286
ConvNext-Base_w8a16.dlc CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6ce78db4eb05531ac3324285f44dd376e0f4aee6764002588927e62af57ee2eb
3
- size 93048948
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:adcecbe0d32abdba5d4bafe930862ddaf3917c56548ab38e53cc4107b1d24de6
3
+ size 93049052
ConvNext-Base_w8a16.onnx.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f3a92ac249ec7f509a871149d3605870f9aee6996895d492af3cb21e2b61c06b
3
- size 303876583
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c0660f41a73ae6647a39e93299abe4d91a991b19162fd7c4e60f175611788bf
3
+ size 303877796
README.md CHANGED
@@ -36,40 +36,42 @@ 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
- | ConvNext-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 41.634 ms | 0 - 231 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
40
- | ConvNext-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 42.29 ms | 1 - 236 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
41
- | ConvNext-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 18.207 ms | 0 - 247 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
42
- | ConvNext-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 20.654 ms | 1 - 249 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
43
- | ConvNext-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 7.327 ms | 0 - 24 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
44
- | ConvNext-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 8.157 ms | 0 - 24 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
45
- | ConvNext-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 7.366 ms | 1 - 23 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
46
- | ConvNext-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 11.39 ms | 0 - 232 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
47
- | ConvNext-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 11.97 ms | 1 - 237 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
48
- | ConvNext-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 5.57 ms | 0 - 244 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
49
- | ConvNext-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 6.1 ms | 1 - 243 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
50
- | ConvNext-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 5.449 ms | 1 - 245 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
51
- | ConvNext-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 4.144 ms | 0 - 233 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
52
- | ConvNext-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 4.602 ms | 1 - 241 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
53
- | ConvNext-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 4.236 ms | 0 - 236 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
54
- | ConvNext-Base | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 3.35 ms | 0 - 235 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
55
- | ConvNext-Base | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 3.531 ms | 1 - 242 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
56
- | ConvNext-Base | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 3.308 ms | 1 - 239 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
57
- | ConvNext-Base | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 8.544 ms | 1254 - 1254 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
58
- | ConvNext-Base | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 7.448 ms | 176 - 176 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
59
- | ConvNext-Base | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 14.412 ms | 0 - 129 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
60
- | ConvNext-Base | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 9.127 ms | 0 - 142 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
61
- | ConvNext-Base | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 5.812 ms | 0 - 32 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
62
- | ConvNext-Base | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 6.092 ms | 0 - 127 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
63
- | ConvNext-Base | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 33.567 ms | 0 - 220 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
64
- | ConvNext-Base | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 607.128 ms | 68 - 88 MB | CPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
65
- | ConvNext-Base | w8a16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 608.474 ms | 86 - 97 MB | CPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
66
- | ConvNext-Base | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 4.16 ms | 0 - 137 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
67
- | ConvNext-Base | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 3.248 ms | 0 - 131 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
68
- | ConvNext-Base | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 261.879 ms | 89 - 119 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
69
- | ConvNext-Base | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 2.505 ms | 0 - 134 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
70
- | ConvNext-Base | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 247.449 ms | 79 - 109 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
71
- | ConvNext-Base | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 6.229 ms | 462 - 462 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
72
- | ConvNext-Base | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 392.38 ms | 133 - 133 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
 
 
73
 
74
 
75
 
@@ -83,9 +85,9 @@ pip install qai-hub-models
83
  ```
84
 
85
 
86
- ## Configure Qualcomm® AI Hub to run this model on a cloud-hosted device
87
 
88
- Sign-in to [Qualcomm® AI Hub](https://app.aihub.qualcomm.com/) with your
89
  Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
90
 
91
  With this API token, you can configure your client to run models on the cloud
@@ -93,7 +95,7 @@ hosted devices.
93
  ```bash
94
  qai-hub configure --api_token API_TOKEN
95
  ```
96
- Navigate to [docs](https://app.aihub.qualcomm.com/docs/) for more information.
97
 
98
 
99
 
@@ -204,7 +206,7 @@ With the output of the model, you can compute like PSNR, relative errors or
204
  spot check the output with expected output.
205
 
206
  **Note**: This on-device profiling and inference requires access to Qualcomm®
207
- AI Hub. [Sign up for access](https://myaccount.qualcomm.com/signup).
208
 
209
 
210
 
 
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
39
+ | ConvNext-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 41.69 ms | 0 - 232 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
40
+ | ConvNext-Base | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 42.358 ms | 1 - 237 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
41
+ | ConvNext-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 18.209 ms | 0 - 246 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
42
+ | ConvNext-Base | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 20.402 ms | 1 - 251 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
43
+ | ConvNext-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 7.405 ms | 0 - 24 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
44
+ | ConvNext-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 8.187 ms | 0 - 24 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
45
+ | ConvNext-Base | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 7.368 ms | 1 - 21 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
46
+ | ConvNext-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 51.498 ms | 0 - 233 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
47
+ | ConvNext-Base | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 53.913 ms | 1 - 238 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
48
+ | ConvNext-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 5.594 ms | 0 - 244 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
49
+ | ConvNext-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 6.055 ms | 1 - 247 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
50
+ | ConvNext-Base | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 5.485 ms | 0 - 249 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
51
+ | ConvNext-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 4.148 ms | 0 - 235 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
52
+ | ConvNext-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 4.599 ms | 1 - 240 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
53
+ | ConvNext-Base | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 4.345 ms | 0 - 237 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
54
+ | ConvNext-Base | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 3.337 ms | 0 - 236 MB | NPU | [ConvNext-Base.tflite](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.tflite) |
55
+ | ConvNext-Base | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 3.551 ms | 1 - 244 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
56
+ | ConvNext-Base | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 3.351 ms | 1 - 240 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
57
+ | ConvNext-Base | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 8.579 ms | 1239 - 1239 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.dlc) |
58
+ | ConvNext-Base | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 7.469 ms | 176 - 176 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base.onnx.zip) |
59
+ | ConvNext-Base | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 14.386 ms | 0 - 131 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
60
+ | ConvNext-Base | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 9.026 ms | 0 - 141 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
61
+ | ConvNext-Base | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 5.834 ms | 0 - 33 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
62
+ | ConvNext-Base | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 6.073 ms | 0 - 129 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
63
+ | ConvNext-Base | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | QNN_DLC | 33.722 ms | 0 - 221 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
64
+ | ConvNext-Base | w8a16 | RB3 Gen 2 (Proxy) | Qualcomm® QCS6490 (Proxy) | ONNX | 612.053 ms | 67 - 87 MB | CPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
65
+ | ConvNext-Base | w8a16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 577.292 ms | 80 - 100 MB | CPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
66
+ | ConvNext-Base | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 4.151 ms | 0 - 140 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
67
+ | ConvNext-Base | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 3.256 ms | 0 - 132 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
68
+ | ConvNext-Base | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 253.107 ms | 89 - 119 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
69
+ | ConvNext-Base | w8a16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 7.759 ms | 0 - 177 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
70
+ | ConvNext-Base | w8a16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | ONNX | 693.199 ms | 69 - 87 MB | CPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
71
+ | ConvNext-Base | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 2.507 ms | 0 - 135 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
72
+ | ConvNext-Base | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 242.051 ms | 78 - 108 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
73
+ | ConvNext-Base | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 6.249 ms | 463 - 463 MB | NPU | [ConvNext-Base.dlc](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.dlc) |
74
+ | ConvNext-Base | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 198.685 ms | 133 - 133 MB | NPU | [ConvNext-Base.onnx.zip](https://huggingface.co/qualcomm/ConvNext-Base/blob/main/ConvNext-Base_w8a16.onnx.zip) |
75
 
76
 
77
 
 
85
  ```
86
 
87
 
88
+ ## Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device
89
 
90
+ Sign-in to [Qualcomm® AI Hub Workbench](https://workbench.aihub.qualcomm.com/) with your
91
  Qualcomm® ID. Once signed in navigate to `Account -> Settings -> API Token`.
92
 
93
  With this API token, you can configure your client to run models on the cloud
 
95
  ```bash
96
  qai-hub configure --api_token API_TOKEN
97
  ```
98
+ Navigate to [docs](https://workbench.aihub.qualcomm.com/docs/) for more information.
99
 
100
 
101
 
 
206
  spot check the output with expected output.
207
 
208
  **Note**: This on-device profiling and inference requires access to Qualcomm®
209
+ AI Hub Workbench. [Sign up for access](https://myaccount.qualcomm.com/signup).
210
 
211
 
212