Hugging Face's logo

qualcomm
/
OpenAI-Clip

Image Classification
PyTorch
LiteRT
foundation
android
Model card Files
xet
Community
1
qaihm-bot commited on
Commit
971feb3
·
verified ·
1 Parent(s): bd4964c

v0.29.1

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.29.1 for changelog.

Files changed (11) hide show
  1. CLIPImageEncoder.bin +0 -3
  2. CLIPImageEncoder.onnx +0 -3
  3. CLIPImageEncoder.so +0 -3
  4. CLIPImageEncoder.tflite +0 -3
  5. CLIPTextEncoder.bin +0 -3
  6. CLIPTextEncoder.onnx +0 -3
  7. CLIPTextEncoder.so +0 -3
  8. CLIPTextEncoder.tflite +0 -3
  9. OpenAI-Clip.bin +0 -3
  10. OpenAI-Clip.so +0 -3
  11. README.md +30 -30
CLIPImageEncoder.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:fd8c0512794a242e67d1280f504b0befeebdfabfda47927ba5cc25e8080c0333
3
- size 176977904
 
 
 
 
CLIPImageEncoder.onnx DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:7ee662d1b6bb5f275ef81cfd4a371b021652a3a1d87cbcdea71ac0074a7380af
3
- size 344918684
 
 
 
 
CLIPImageEncoder.so DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:fb2dc67941ffa39cd03c7dc54259bb7aa69298e6563f7ff82be9e4789765fd82
3
- size 345882696
 
 
 
 
CLIPImageEncoder.tflite DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:cf189e60bbbe3d9bae052b9679ae5f09678475f639c6e7092d2ce0bf0ce1a9a4
3
- size 344886032
 
 
 
 
CLIPTextEncoder.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:8be3e617f82ac57d51d926c48c8597f97832bb114c76f8cbb931552f1f4625d3
3
- size 128515512
 
 
 
 
CLIPTextEncoder.onnx DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:8dc504d03bea97e1d04ceff0b99fb550fc0f67dc5d2ae5e2d7e2ecdcdf71eaed
3
- size 253870537
 
 
 
 
CLIPTextEncoder.so DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:83cfcef3950c9bab2a760d43cfa2cec6c9ea3d197d8d0e049436e8a86ae59d6f
3
- size 254820312
 
 
 
 
CLIPTextEncoder.tflite DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:b01a335e80e2f02ccdf28fa7076085d809faa12236da53c04ed9222cbd195fcc
3
- size 253847848
 
 
 
 
OpenAI-Clip.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:fdb4400f077f56e7111a508fa5e9e880349a74314f9cf365de067028ff97ae29
3
- size 305632512
 
 
 
 
OpenAI-Clip.so DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:9a06c6a9319bd3e21e44396e7e5024ec398d80d0b70324678fe38248dc5cd1b6
3
- size 600347328
 
 
 
 
README.md CHANGED
@@ -38,35 +38,35 @@ More details on model performance across various devices, can be found
38
 
39
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
40
  |---|---|---|---|---|---|---|---|---|
41
- | OpenAI-Clip | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 73.04 ms | 0 - 382 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
42
- | OpenAI-Clip | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN | 60.196 ms | 1 - 10 MB | NPU | Use Export Script |
43
- | OpenAI-Clip | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 27.852 ms | 0 - 345 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
44
- | OpenAI-Clip | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN | 24.464 ms | 0 - 414 MB | NPU | Use Export Script |
45
- | OpenAI-Clip | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 24.863 ms | 0 - 21 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
46
- | OpenAI-Clip | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN | 20.977 ms | 1 - 3 MB | NPU | Use Export Script |
47
- | OpenAI-Clip | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 29.595 ms | 0 - 381 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
48
- | OpenAI-Clip | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN | 23.856 ms | 0 - 10 MB | NPU | Use Export Script |
49
- | OpenAI-Clip | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 73.04 ms | 0 - 382 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
50
- | OpenAI-Clip | float | SA7255P ADP | Qualcomm® SA7255P | QNN | 60.196 ms | 1 - 10 MB | NPU | Use Export Script |
51
- | OpenAI-Clip | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 24.952 ms | 0 - 26 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
52
- | OpenAI-Clip | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN | 21.045 ms | 1 - 2 MB | NPU | Use Export Script |
53
- | OpenAI-Clip | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 30.37 ms | 0 - 336 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
54
- | OpenAI-Clip | float | SA8295P ADP | Qualcomm® SA8295P | QNN | 24.945 ms | 1 - 17 MB | NPU | Use Export Script |
55
- | OpenAI-Clip | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 24.941 ms | 0 - 21 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
56
- | OpenAI-Clip | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN | 21.002 ms | 1 - 3 MB | NPU | Use Export Script |
57
- | OpenAI-Clip | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 29.595 ms | 0 - 381 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
58
- | OpenAI-Clip | float | SA8775P ADP | Qualcomm® SA8775P | QNN | 23.856 ms | 0 - 10 MB | NPU | Use Export Script |
59
- | OpenAI-Clip | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 24.949 ms | 0 - 25 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
60
- | OpenAI-Clip | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN | 20.902 ms | 0 - 47 MB | NPU | Use Export Script |
61
- | OpenAI-Clip | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 25.496 ms | 0 - 243 MB | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.onnx) |
62
- | OpenAI-Clip | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 17.882 ms | 0 - 382 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
63
- | OpenAI-Clip | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN | 15.184 ms | 0 - 417 MB | NPU | Use Export Script |
64
- | OpenAI-Clip | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 18.235 ms | 1 - 474 MB | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.onnx) |
65
- | OpenAI-Clip | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 14.681 ms | 0 - 382 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
66
- | OpenAI-Clip | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN | 13.878 ms | 1 - 388 MB | NPU | Use Export Script |
67
- | OpenAI-Clip | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 17.205 ms | 1 - 443 MB | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.onnx) |
68
- | OpenAI-Clip | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 21.687 ms | 1 - 1 MB | NPU | Use Export Script |
69
- | OpenAI-Clip | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 26.597 ms | 293 - 293 MB | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.onnx) |
70
 
71
 
72
 
@@ -130,7 +130,7 @@ Profiling Results
130
  OpenAI-Clip
131
  Device : cs_8275 (ANDROID 14)
132
  Runtime : TFLITE
133
- Estimated inference time (ms) : 73.0
134
  Estimated peak memory usage (MB): [0, 382]
135
  Total # Ops : 1320
136
  Compute Unit(s) : npu (1318 ops) gpu (0 ops) cpu (2 ops)
 
38
 
39
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
40
  |---|---|---|---|---|---|---|---|---|
41
+ | OpenAI-Clip | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 368.708 ms | 0 - 382 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
42
+ | OpenAI-Clip | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN | 60.22 ms | 1 - 10 MB | NPU | Use Export Script |
43
+ | OpenAI-Clip | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 27.434 ms | 0 - 348 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
44
+ | OpenAI-Clip | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN | 24.42 ms | 1 - 414 MB | NPU | Use Export Script |
45
+ | OpenAI-Clip | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 24.947 ms | 0 - 22 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
46
+ | OpenAI-Clip | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN | 21.013 ms | 1 - 3 MB | NPU | Use Export Script |
47
+ | OpenAI-Clip | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 29.716 ms | 0 - 381 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
48
+ | OpenAI-Clip | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN | 23.796 ms | 0 - 10 MB | NPU | Use Export Script |
49
+ | OpenAI-Clip | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 368.708 ms | 0 - 382 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
50
+ | OpenAI-Clip | float | SA7255P ADP | Qualcomm® SA7255P | QNN | 60.22 ms | 1 - 10 MB | NPU | Use Export Script |
51
+ | OpenAI-Clip | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 25.171 ms | 0 - 21 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
52
+ | OpenAI-Clip | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN | 21.012 ms | 1 - 3 MB | NPU | Use Export Script |
53
+ | OpenAI-Clip | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 30.38 ms | 0 - 336 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
54
+ | OpenAI-Clip | float | SA8295P ADP | Qualcomm® SA8295P | QNN | 24.984 ms | 1 - 17 MB | NPU | Use Export Script |
55
+ | OpenAI-Clip | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 25.15 ms | 0 - 21 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
56
+ | OpenAI-Clip | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN | 21.04 ms | 1 - 2 MB | NPU | Use Export Script |
57
+ | OpenAI-Clip | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 29.716 ms | 0 - 381 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
58
+ | OpenAI-Clip | float | SA8775P ADP | Qualcomm® SA8775P | QNN | 23.796 ms | 0 - 10 MB | NPU | Use Export Script |
59
+ | OpenAI-Clip | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 25.295 ms | 0 - 21 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
60
+ | OpenAI-Clip | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN | 21.074 ms | 0 - 55 MB | NPU | Use Export Script |
61
+ | OpenAI-Clip | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 25.33 ms | 0 - 312 MB | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.onnx) |
62
+ | OpenAI-Clip | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 17.728 ms | 0 - 383 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
63
+ | OpenAI-Clip | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN | 15.132 ms | 14 - 431 MB | NPU | Use Export Script |
64
+ | OpenAI-Clip | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 18.105 ms | 1 - 475 MB | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.onnx) |
65
+ | OpenAI-Clip | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 16.906 ms | 0 - 382 MB | NPU | [OpenAI-Clip.tflite](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.tflite) |
66
+ | OpenAI-Clip | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN | 11.865 ms | 1 - 390 MB | NPU | Use Export Script |
67
+ | OpenAI-Clip | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 17.747 ms | 1 - 442 MB | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.onnx) |
68
+ | OpenAI-Clip | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 21.682 ms | 1 - 1 MB | NPU | Use Export Script |
69
+ | OpenAI-Clip | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 26.59 ms | 293 - 293 MB | NPU | [OpenAI-Clip.onnx](https://huggingface.co/qualcomm/OpenAI-Clip/blob/main/OpenAI-Clip.onnx) |
70
 
71
 
72
 
 
130
  OpenAI-Clip
131
  Device : cs_8275 (ANDROID 14)
132
  Runtime : TFLITE
133
+ Estimated inference time (ms) : 368.7
134
  Estimated peak memory usage (MB): [0, 382]
135
  Total # Ops : 1320
136
  Compute Unit(s) : npu (1318 ops) gpu (0 ops) cpu (2 ops)