-
Notifications
You must be signed in to change notification settings - Fork 257
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Update Examples for TF 3x API (#1901)
Signed-off-by: zehao-intel <zehao.huang@intel.com>
- Loading branch information
1 parent
6b30207
commit fb85779
Showing
117 changed files
with
39,154 additions
and
31 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,18 @@ | ||
| { | ||
| "keras": { | ||
| "resnetv2_50": { | ||
| "model_src_dir": "keras/image_recognition/resnet_v2_50/quantization/ptq", | ||
| "dataset_location": "/tf_dataset/dataset/imagenet", | ||
| "input_model": "/tf_dataset2/models/tensorflow/resnetv2_50_keras/saved_model", | ||
| "main_script": "main.py", | ||
| "batch_size": 32 | ||
| }, | ||
| "inception_v3": { | ||
| "model_src_dir": "keras/image_recognition/inception_v3/quantization/ptq", | ||
| "dataset_location": "/tf_dataset/dataset/imagenet", | ||
| "input_model": "/tf_dataset2/models/tensorflow/inception_v3_keras/saved_model", | ||
| "main_script": "main.py", | ||
| "batch_size": 32 | ||
| } | ||
| } | ||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
75 changes: 75 additions & 0 deletions
75
...es/3.x_api/tensorflow/image_recognition/inception_v3/quantization/ptq/README.md
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,75 @@ | ||
| Step-by-Step | ||
| ============ | ||
|
|
||
| This document list steps of reproducing inception_v3 model tuning and benchmark results via Neural Compressor. | ||
| This example can run on Intel CPUs and GPUs. | ||
|
|
||
| > **Note**: | ||
| > The models is supported in validated TensorFlow [Version](/docs/source/installation_guide.md#validated-software-environment). | ||
| # Prerequisite | ||
|
|
||
| ## 1. Environment | ||
|
|
||
| ### Installation | ||
| Recommend python 3.9 or higher version. | ||
| ```shell | ||
| pip install -r requirements.txt | ||
| ``` | ||
|
|
||
| ### Install Intel Extension for Tensorflow | ||
| #### Quantizing the model on Intel GPU(Mandatory to install ITEX) | ||
| Intel Extension for Tensorflow is mandatory to be installed for quantizing the model on Intel GPUs. | ||
|
|
||
| ```shell | ||
| pip install --upgrade intel-extension-for-tensorflow[xpu] | ||
| ``` | ||
| For any more details, please follow the procedure in [install-gpu-drivers](https://github.com/intel/intel-extension-for-tensorflow/blob/main/docs/install/install_for_xpu.md#install-gpu-drivers) | ||
|
|
||
| #### Quantizing the model on Intel CPU(Optional to install ITEX) | ||
| Intel Extension for Tensorflow for Intel CPUs is experimental currently. It's not mandatory for quantizing the model on Intel CPUs. | ||
|
|
||
| ```shell | ||
| pip install --upgrade intel-extension-for-tensorflow[cpu] | ||
| ``` | ||
| > **Note**: | ||
| > The version compatibility of stock Tensorflow and ITEX can be checked [here](https://github.com/intel/intel-extension-for-tensorflow#compatibility-table). Please make sure you have installed compatible Tensorflow and ITEX. | ||
| ## 2. Prepare pre-trained model | ||
|
|
||
| Download pre-trained PB | ||
| ```shell | ||
| wget https://storage.googleapis.com/intel-optimized-tensorflow/models/v1_6/inceptionv3_fp32_pretrained_model.pb | ||
| ``` | ||
|
|
||
| ## 3. Prepare Dataset | ||
|
|
||
| TensorFlow [models](https://github.com/tensorflow/models) repo provides [scripts and instructions](https://github.com/tensorflow/models/tree/master/research/slim#an-automated-script-for-processing-imagenet-data) to download, process and convert the ImageNet dataset to the TF records format. | ||
| We also prepared related scripts in ` examples/3.x_api/tensorflow/cv` directory. To download the raw images, the user must create an account with image-net.org. If you have downloaded the raw data and preprocessed the validation data by moving the images into the appropriate sub-directory based on the label (synset) of the image. we can use below command ro convert it to tf records format. | ||
|
|
||
| ```shell | ||
| cd examples/3.x_api/tensorflow/cv | ||
| # convert validation subset | ||
| bash prepare_dataset.sh --output_dir=./inception_v3/quantization/ptq/data --raw_dir=/PATH/TO/img_raw/val/ --subset=validation | ||
| # convert train subset | ||
| bash prepare_dataset.sh --output_dir=./inception_v3/quantization/ptq/data --raw_dir=/PATH/TO/img_raw/train/ --subset=train | ||
| ``` | ||
| > **Note**: | ||
| > The raw ImageNet dataset resides in JPEG files should be in the following directory structure. Taking validation set as an example:<br> | ||
| > /PATH/TO/img_raw/val/n01440764/ILSVRC2012_val_00000293.JPEG<br> | ||
| > /PATH/TO/img_raw/val/n01440764/ILSVRC2012_val_00000543.JPEG<br> | ||
| > where 'n01440764' is the unique synset label associated with these images. | ||
|
|
||
| # Run | ||
|
|
||
| ## 1 Quantization | ||
|
|
||
| ```shell | ||
| bash run_quant.sh --input_model=/PATH/TO/inceptionv3_fp32_pretrained_model.pb \ | ||
| --output_model=./nc_inception_v3.pb --dataset_location=/path/to/ImageNet/ | ||
| ``` | ||
|
|
||
| ## 2. Benchmark | ||
| ```shell | ||
| bash run_benchmark.sh --input_model=./nc_inception_v3.pb --mode=accuracy --dataset_location=/path/to/ImageNet/ --batch_size=32 | ||
| bash run_benchmark.sh --input_model=./nc_inception_v3.pb --mode=performance --dataset_location=/path/to/ImageNet/ --batch_size=1 | ||
| ``` |
Oops, something went wrong.