Skip to content

Update onnx example for new API #190

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 13 commits into from
Dec 29, 2022
Merged

Update onnx example for new API #190

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
a91ef69
update example with new API
mengniwang95 Dec 9, 2022
f194b14
update link and script
mengniwang95 Dec 9, 2022
59380b4
add bs
mengniwang95 Dec 15, 2022
b1142fc
rebase and fix bug
mengniwang95 Dec 19, 2022
59d9c85
fix bug
mengniwang95 Dec 21, 2022
4790772
fix config
mengniwang95 Dec 23, 2022
687c3a4
fix script
mengniwang95 Dec 26, 2022
d272a50
fix fcn
mengniwang95 Dec 26, 2022
6899319
fix bug
mengniwang95 Dec 27, 2022
92deb8d
remove paramter
mengniwang95 Dec 27, 2022
d1fbf0f
fix perf
mengniwang95 Dec 28, 2022
4560721
remove example
mengniwang95 Dec 29, 2022
104ce6b
update config
mengniwang95 Dec 29, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
44 changes: 16 additions & 28 deletions examples/.config/model_params_onnxrt.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -182,12 +182,10 @@
},
"squeezenet": {
"model_src_dir": "image_recognition/onnx_model_zoo/squeezenet/quantization/ptq",
"dataset_location": "/tf_dataset2/datasets/imagenet/ImagenetRaw/ILSVRC2012_img_val",
"dataset_location": "/tf_dataset2/datasets/imagenet/ImagenetRaw",
"input_model": "/tf_dataset2/models/onnx/squeezenet/squeezenet1.0-12.onnx",
"yaml": "squeezenet.yaml",
"strategy": "basic",
"batch_size": 1,
"new_benchmark": true
"main_script": "main.py",
"batch_size": 1
},
"caffenet": {
"model_src_dir": "image_recognition/onnx_model_zoo/caffenet/quantization/ptq",
Expand Down Expand Up @@ -245,12 +243,10 @@
},
"fcn": {
"model_src_dir": "image_recognition/onnx_model_zoo/fcn/quantization/ptq",
"dataset_location": "/tf_dataset2/datasets/coco2017/coco/val2017",
"dataset_location": "/tf_dataset2/datasets/coco2017/coco",
"input_model": "/tf_dataset2/models/onnx/fcn/fcn-resnet50-11.onnx",
"yaml": "fcn_rn50.yaml",
"strategy": "basic",
"batch_size": 1,
"new_benchmark": true
"main_script": "main.py",
"batch_size": 1
},
"ssd_mobilenet_v1-2": {
"model_src_dir": "object_detection/onnx_model_zoo/ssd_mobilenet_v1/quantization/ptq",
Expand Down Expand Up @@ -533,12 +529,10 @@
},
"fcn_qdq": {
"model_src_dir": "image_recognition/onnx_model_zoo/fcn/quantization/ptq",
"dataset_location": "/tf_dataset2/datasets/coco2017/coco/val2017",
"dataset_location": "/tf_dataset2/datasets/coco2017/coco",
"input_model": "/tf_dataset2/models/onnx/fcn/fcn-resnet50-13.onnx",
"yaml": "fcn_rn50_qdq.yaml",
"strategy": "basic",
"batch_size": 1,
"new_benchmark": true
"main_script": "main.py",
"batch_size": 1
},
"ssd_mobilenet_v1-2_qdq": {
"model_src_dir": "object_detection/onnx_model_zoo/ssd_mobilenet_v1/quantization/ptq",
Expand Down Expand Up @@ -595,31 +589,25 @@
"new_benchmark": true
},
"arcface": {
"model_src_dir": "image_recognition/onnx_model_zoo/arcface/quantization/ptq",
"model_src_dir": "body_analysis/onnx_model_zoo/arcface/quantization/ptq",
"dataset_location": "/tf_dataset2/datasets/faces_ms1m_112x112/lfw.bin",
"input_model": "/tf_dataset2/models/onnx/arcface/arcfaceresnet100-11.onnx",
"yaml": "arcface.yaml",
"strategy": "basic",
"batch_size": 1,
"new_benchmark": true
"main_script": "main.py",
"batch_size": 1
},
"emotion_ferplus": {
"model_src_dir": "body_analysis/onnx_model_zoo/emotion_ferplus/quantization/ptq",
"dataset_location": "/tf_dataset2/datasets/emotion_data/emotion_ferplus/fer2013/fer2013.csv",
"input_model": "/tf_dataset2/models/onnx/emotion_ferplus/emotion-ferplus-12.onnx",
"yaml": "emotion_ferplus.yaml",
"strategy": "basic",
"batch_size": 1,
"new_benchmark": true
"main_script": "main.py",
"batch_size": 1
},
"ultraface": {
"model_src_dir": "body_analysis/onnx_model_zoo/ultraface/quantization/ptq",
"dataset_location": "/tf_dataset2/datasets/wider_face_data/WIDER_val",
"input_model": "/tf_dataset2/models/onnx/ultraface/version-RFB-320-12.onnx",
"yaml": "ultraface.yaml",
"strategy": "basic",
"batch_size": 1,
"new_benchmark": true
"main_script": "main.py",
"batch_size": 1
},
"unet": {
"model_src_dir": "image_recognition/unet/quantization/ptq",
Expand Down
38 changes: 19 additions & 19 deletions examples/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -774,55 +774,55 @@ Intel® Neural Compressor validated examples with multiple compression technique
<td>ResNet50 V1.5</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/resnet50/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/resnet50/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/resnet50/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/resnet50/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>ResNet50 V1.5 MLPerf</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/resnet50/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/resnet50/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/resnet50/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/resnet50/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>VGG16</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/vgg16/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/vgg16/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/vgg16/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/vgg16/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>MobileNet V2</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/mobilenet_v2/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/mobilenet_v2/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/mobilenet_v2/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/mobilenet_v2/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>MobileNet V3 MLPerf</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/mobilenet_v3/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/mobilenet_v3/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/mobilenet_v3/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/mobilenet_v3/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>AlexNet</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/alexnet/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/alexnet/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/alexnet/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/alexnet/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>CaffeNet</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/caffenet/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/caffenet/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/caffenet/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/caffenet/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>DenseNet</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/densenet/quantization/ptq">qlinearops</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/densenet/quantization/ptq">qlinearops</a></td>
</tr>
<tr>
<td>EfficientNet</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/efficientnet/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/efficientnet/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/efficientnet/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/efficientnet/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>FCN</td>
Expand All @@ -834,37 +834,37 @@ Intel® Neural Compressor validated examples with multiple compression technique
<td>GoogleNet</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/googlenet/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/googlenet/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/googlenet/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/googlenet/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>Inception V1</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/inception/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/inception/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/inception/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/inception/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>MNIST</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/mnist/quantization/ptq">qlinearops</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/mnist/quantization/ptq">qlinearops</a></td>
</tr>
<tr>
<td>MobileNet V2 (ONNX Model Zoo)</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/mobilenet/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/mobilenet/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/mobilenet/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/mobilenet/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>ResNet50 V1.5 (ONNX Model Zoo)</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/resnet50/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/resnet50/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/resnet50/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/resnet50/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>ShuffleNet V2</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/shufflenet/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/shufflenet/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/shufflenet/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/shufflenet/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>SqueezeNet</td>
Expand All @@ -876,19 +876,19 @@ Intel® Neural Compressor validated examples with multiple compression technique
<td>VGG16 (ONNX Model Zoo)</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/vgg16/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/vgg16/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/vgg16/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/vgg16/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>ZFNet</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/zfnet/quantization/ptq">qlinearops</a> / <a href="./onnxrt/image_recognition/onnx_model_zoo/zfnet/quantization/ptq">qdq</a></td>
<td><a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/zfnet/quantization/ptq">qlinearops</a> / <a href="https://github.com/intel/neural-compressor/tree/old_api_examples/examples/onnxrt/image_recognition/onnx_model_zoo/zfnet/quantization/ptq">qdq</a></td>
</tr>
<tr>
<td>ArcFace</td>
<td>Image Recognition</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/image_recognition/onnx_model_zoo/arcface/quantization/ptq">qlinearops</a></td>
<td><a href="./onnxrt/body_analysis/onnx_model_zoo/arcface/quantization/ptq">qlinearops</a></td>
</tr>
<tr>
<td>BERT base MRPC</td>
Expand Down Expand Up @@ -1068,7 +1068,7 @@ Intel® Neural Compressor validated examples with multiple compression technique
<td>Emotion FERPlus</td>
<td>Body Analysis</td>
<td>Post-Training Static Quantization</td>
<td><a href="./onnxrt/body_analysis/onnx_model_zoo/emotion_ferplus/quantization/ptq">qlinearops</a></td>
<td><a href="/onnxrt/body_analysis/onnx_model_zoo/emotion_ferplus/quantization/ptq">qlinearops</a></td>
</tr>
<tr>
<td>Ultra Face</td>
Expand Down
8 changes: 2 additions & 6 deletions ...el_zoo/arcface/quantization/ptq/README.md → ...el_zoo/arcface/quantization/ptq/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,19 +18,15 @@ wget https://github.com/onnx/models/raw/main/vision/body_analysis/arcface/model/

```bash
bash run_tuning.sh --input_model=path/to/model \ # model path as *.onnx
--config=arcface.yaml \
--data_path=/path/to/faces_ms1m_112x112/task.bin \
--nfolds=nfolds_num \ # optional, data type is int, default is 1
--dataset_location=/path/to/faces_ms1m_112x112/task.bin \
--output_model=path/to/save
```

### Benchmark

```bash
bash run_benchmark.sh --input_model=path/to/model \ # model path as *.onnx
--config=arcface.yaml \
--data_path=/path/to/faces_ms1m_112x112/task.bin \
--nfolds=nfolds_num \ # optional, data type is int, default is 1
--dataset_location=/path/to/faces_ms1m_112x112/task.bin \
--mode=performance # or accuracy
```

Loading