Hugging Face's logo

qualcomm
/
TrOCR

Image-to-Text
PyTorch
LiteRT
android
Model card Files Community
3
qaihm-bot commited on
Commit
c651c25
·
verified ·
1 Parent(s): 4dff859

v0.32.0

Browse files

See https://github.com/quic/ai-hub-models/releases/v0.32.0 for changelog.

Files changed (6) hide show
  1. .gitattributes +1 -0
  2. DEPLOYMENT_MODEL_LICENSE.pdf +3 -0
  3. LICENSE +2 -0
  4. README.md +11 -11
  5. TrOCR_TrOCRDecoder.onnx +2 -2
  6. TrOCR_TrOCREncoder.onnx +2 -2
.gitattributes CHANGED
@@ -37,3 +37,4 @@ TrOCREncoder.so filter=lfs diff=lfs merge=lfs -text
37
  TrOCRDecoder.so filter=lfs diff=lfs merge=lfs -text
38
  TrOCR_TrOCRDecoder.dlc filter=lfs diff=lfs merge=lfs -text
39
  TrOCR_TrOCREncoder.dlc filter=lfs diff=lfs merge=lfs -text
 
 
37
  TrOCRDecoder.so filter=lfs diff=lfs merge=lfs -text
38
  TrOCR_TrOCRDecoder.dlc filter=lfs diff=lfs merge=lfs -text
39
  TrOCR_TrOCREncoder.dlc filter=lfs diff=lfs merge=lfs -text
40
+ DEPLOYMENT_MODEL_LICENSE.pdf filter=lfs diff=lfs merge=lfs -text
DEPLOYMENT_MODEL_LICENSE.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4409f93b0e82531303b3e10f52f1fdfb56467a25f05b7441c6bbd8bb8a64b42c
3
+ size 109629
LICENSE ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ The license of the original trained model can be found at https://github.com/microsoft/unilm/blob/master/LICENSE.
2
+ The license for the deployable model files (.tflite, .onnx, .dlc, .bin, etc.) can be found in DEPLOYMENT_MODEL_LICENSE.pdf.
README.md CHANGED
@@ -29,10 +29,10 @@ More details on model performance across various devices, can be found
29
  - **Model Stats:**
30
  - Model checkpoint: trocr-small-stage1
31
  - Input resolution: 320x320
32
- - Number of parameters (TrOCREncoder): 23.0M
33
- - Model size (TrOCREncoder): 87.8 MB
34
  - Number of parameters (TrOCRDecoder): 38.3M
35
- - Model size (TrOCRDecoder): 146 MB
 
 
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
@@ -56,15 +56,15 @@ More details on model performance across various devices, can be found
56
  | TrOCRDecoder | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 2.809 ms | 0 - 64 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
57
  | TrOCRDecoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 2.07 ms | 0 - 515 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
58
  | TrOCRDecoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 2.034 ms | 2 - 27 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
59
- | TrOCRDecoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 2.61 ms | 0 - 243 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
60
  | TrOCRDecoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.497 ms | 0 - 156 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
61
  | TrOCRDecoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.499 ms | 0 - 144 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
62
- | TrOCRDecoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.059 ms | 0 - 145 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
63
  | TrOCRDecoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 1.366 ms | 0 - 79 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
64
  | TrOCRDecoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 1.357 ms | 1 - 155 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
65
- | TrOCRDecoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.675 ms | 1 - 157 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
66
  | TrOCRDecoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 2.279 ms | 675 - 675 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
67
- | TrOCRDecoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.371 ms | 68 - 68 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
68
  | TrOCREncoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 77.528 ms | 7 - 167 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
69
  | TrOCREncoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 72.833 ms | 2 - 149 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
70
  | TrOCREncoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 52.189 ms | 7 - 171 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
@@ -85,15 +85,15 @@ More details on model performance across various devices, can be found
85
  | TrOCREncoder | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 39.725 ms | 2 - 150 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
86
  | TrOCREncoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 40.941 ms | 7 - 31 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
87
  | TrOCREncoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 38.753 ms | 2 - 37 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
88
- | TrOCREncoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 40.111 ms | 14 - 140 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
89
  | TrOCREncoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 32.336 ms | 134 - 296 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
90
  | TrOCREncoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 31.86 ms | 2 - 155 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
91
- | TrOCREncoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 32.19 ms | 12 - 169 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
92
  | TrOCREncoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 28.294 ms | 5 - 165 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
93
  | TrOCREncoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 28.687 ms | 2 - 176 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
94
- | TrOCREncoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 27.961 ms | 14 - 203 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
95
  | TrOCREncoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 37.201 ms | 102 - 102 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
96
- | TrOCREncoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 37.174 ms | 50 - 50 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
97
 
98
 
99
 
 
29
  - **Model Stats:**
30
  - Model checkpoint: trocr-small-stage1
31
  - Input resolution: 320x320
 
 
32
  - Number of parameters (TrOCRDecoder): 38.3M
33
+ - Model size (TrOCRDecoder) (float): 146 MB
34
+ - Number of parameters (TrOCREncoder): 23.0M
35
+ - Model size (TrOCREncoder) (float): 87.8 MB
36
 
37
  | Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
38
  |---|---|---|---|---|---|---|---|---|
 
56
  | TrOCRDecoder | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 2.809 ms | 0 - 64 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
57
  | TrOCRDecoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 2.07 ms | 0 - 515 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
58
  | TrOCRDecoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 2.034 ms | 2 - 27 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
59
+ | TrOCRDecoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 2.497 ms | 0 - 223 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
60
  | TrOCRDecoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 1.497 ms | 0 - 156 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
61
  | TrOCRDecoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.499 ms | 0 - 144 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
62
+ | TrOCRDecoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 1.899 ms | 0 - 142 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
63
  | TrOCRDecoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 1.366 ms | 0 - 79 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
64
  | TrOCRDecoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 1.357 ms | 1 - 155 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
65
+ | TrOCRDecoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 1.673 ms | 2 - 158 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
66
  | TrOCRDecoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 2.279 ms | 675 - 675 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
67
+ | TrOCRDecoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 2.352 ms | 68 - 68 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
68
  | TrOCREncoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 77.528 ms | 7 - 167 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
69
  | TrOCREncoder | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 72.833 ms | 2 - 149 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
70
  | TrOCREncoder | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 52.189 ms | 7 - 171 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
 
85
  | TrOCREncoder | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 39.725 ms | 2 - 150 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
86
  | TrOCREncoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | TFLITE | 40.941 ms | 7 - 31 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
87
  | TrOCREncoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | QNN_DLC | 38.753 ms | 2 - 37 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
88
+ | TrOCREncoder | float | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 Mobile | ONNX | 39.704 ms | 7 - 134 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
89
  | TrOCREncoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 32.336 ms | 134 - 296 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
90
  | TrOCREncoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 31.86 ms | 2 - 155 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
91
+ | TrOCREncoder | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 32.373 ms | 12 - 170 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
92
  | TrOCREncoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | TFLITE | 28.294 ms | 5 - 165 MB | NPU | [TrOCR.tflite](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.tflite) |
93
  | TrOCREncoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | QNN_DLC | 28.687 ms | 2 - 176 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
94
+ | TrOCREncoder | float | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite Mobile | ONNX | 24.852 ms | 14 - 203 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
95
  | TrOCREncoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 37.201 ms | 102 - 102 MB | NPU | [TrOCR.dlc](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.dlc) |
96
+ | TrOCREncoder | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 37.106 ms | 50 - 50 MB | NPU | [TrOCR.onnx](https://huggingface.co/qualcomm/TrOCR/blob/main/TrOCR.onnx) |
97
 
98
 
99
 
TrOCR_TrOCRDecoder.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:821abf014bbd55cf0c705010b17d9780080d8568f652a3b8f6a38735386756cd
3
- size 153400081
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb6e991751816bbb1a330d7efbd9dd44b4607497914ce16b36bd26c66fa7f7b5
3
+ size 153408540
TrOCR_TrOCREncoder.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c5d88cf9fdf0e58cb8659265bdfc1d03a7e3e2487729410c5861da190a1474ed
3
- size 92105049
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfdb3bcd2b7f30a8e1bb7fe7941dbfd5b307c2a06d92195ae0ebf7dce99c7bff
3
+ size 92116702