Skip to content

Latest commit

 

History

History
141 lines (111 loc) · 7.63 KB

README.md

File metadata and controls

141 lines (111 loc) · 7.63 KB
Raw

Instance Segmentation Sample

This sample demonstrates DL model compression capabilities for Instance Segmentation task. The sample consists of basic steps such as DL model initialization, dataset preparation, training loop over epochs and validation steps. The sample receives a configuration file where the training schedule, hyper-parameters, and compression settings are defined.

Features

  • Mask R-CNN from the official TF repository with minor modifications (custom implementation of upsamling is replaced with equivalent tf.keras.layers.UpSampling2D).
  • Support TFRecords for COCO2017 dataset.
  • Configuration file examples for sparsity, quantization, and quantization with sparsity.
  • Export to Frozen Graph or TensorFlow SavedModel that is supported by the OpenVINO™ toolkit.
  • Distributed training on multiple GPUs on one machine is supported using tf.distribute.MirroredStrategy.

Installation

To work with the sample you should install the corresponding Python package dependencies

pip install -r examples/tensorflow/requirements.txt

Quantize Pretrained Model

This scenario demonstrates quantization with fine-tuning of Mask R-CNN with ResNet-50 backbone on the COCO2017 dataset.

Dataset Preparation

The instance segmentation sample supports dataset only in TFRecords format.

To download the COCO2017 dataset and convert it to TFRecord format please use download_and_preprocess_coco.sh script from the official TensorFlow TPU repository.

bash <path_to_tensorflow_tpu_repo>/tools/datasets/download_and_preprocess_coco.sh <path_to_coco_data_dir>

This script installs the required libraries and then runs the dataset preprocessing. The output of the script is *.tfrecord files in your local data directory.

The COCO2017 dataset should be specified in the configuration file as follows:

"dataset": "coco/2017"

Run Instance Segmentation Sample

We can run the sample after data preparation. For this follow these steps:

  • If you did not install the package, add the repository root folder to the PYTHONPATH environment variable.
  • Go to the examples/tensorflow/segmentation folder.
  • Download the pre-trained weights in checkpoint format and provide the path to them using --weights flag. The link to the archive with pre-trained weights can be found in the TensorFlow checkpoint column of the results table. Select the checkpoint corresponding to the None compression algorithm, which includes the pre-trained weights for the FP32 model, without applying any compression algorithms.
  • Specify the GPUs to be used for training by setting the environment variable CUDA_VISIBLE_DEVICES. This is necessary because training and validation during training must be performed on different GPU devices. Please note that usually only one GPU is required for validation during training.
  • Run the following command to start compression with fine-tuning on all available GPUs on the machine: 
    python train.py \
--config=configs/quantization/mask_rcnn_coco_int8.json \
--weights=<path_to_ckpt_file_with_pretrained_weights> \
--data=<path_to_dataset> \
--log-dir=../../results/quantization/maskrcnn_coco_int8
  • Use the --resume flag with the path to the checkpoint to resume training from the defined checkpoint or folder with checkpoints to resume training from the last checkpoint.

To start checkpoints validation during training follow these steps:

  • If you did not install the package, add the repository root folder to the PYTHONPATH environment variable.
  • Go to the examples/tensorflow/segmentation folder.
  • Specify the GPUs to be used for validation during training by setting the environment variable CUDA_VISIBLE_DEVICES.
  • Run the following command to start checkpoints validation during training: 
    python evaluation.py \
--mode=train \
--config=configs/quantization/mask_rcnn_coco_int8.json \
--data=<path_to_dataset> \
--batch-size=1 \
--checkpoint-save-dir=<path_to_checkpoints>

Validate Your Model Checkpoint

To estimate the test scores of your model checkpoint, use the following command

python evaluation.py \
--mode=test \
--config=configs/quantization/mask_rcnn_coco_int8.json \
--data=<path_to_dataset> \
--batch-size=1 \
--resume=<path_to_trained_model_checkpoint>

To validate an model checkpoint, make sure the compression algorithm settings are empty in the configuration file and path to checkpoint file with model weights is provided in command line argument --weights

Export Compressed Model

To export trained model to the Frozen Graph, use the following command:

python evaluation.py \
--mode=export \
--config=configs/quantization/mask_rcnn_coco_int8.json \
--batch-size=1 \
--resume=<path_to_trained_model_checkpoint> \
--to-frozen-graph=../../results/mask_rcnn_coco_int8.pb

To export trained model to the SavedModel, use the following command:

python evaluation.py \
--mode=export \
--config=configs/quantization/mask_rcnn_coco_int8.json \
--batch-size=1 \
--resume=<path_to_trained_model_checkpoint> \
--to-saved-model=../../results/saved_model

Export to OpenVINO™ Intermediate Representation (IR)

To export a model to the OpenVINO IR and run it using the Intel® Deep Learning Deployment Toolkit, refer to this tutorial.

Train MaskRCNN from scratch

  • Download pre-trained ResNet-50 checkpoint from here.
  • If you did not install the package, add the repository root folder to the PYTHONPATH environment variable.
  • Go to the examples/tensorflow/segmentation folder.
  • Run the following command to start training MaskRCNN from scratch on all available GPUs on the machine: 
    python train.py \
--config=configs/mask_rcnn_coco.json \
--backbone-checkpoint=<path_to_resnet50-2018-02-07_folder> \
--data=<path_to_dataset> \
--log-dir=../../results/quantization/maskrcnn_coco_baseline

Results

Model Compression algorithm Dataset mAP (drop) % NNCF config file TensorFlow checkpoint
MaskRCNN None COCO2017 bbox: 37.33
segm: 33.56		 Link Link
MaskRCNN INT8 (per-tensor, symmetric for weights; per-tensor, symmetric for activations) COCO2017 bbox: 37.14 (0.19)
segm: 33.53 (0.03)		 Link Link
MaskRCNN Sparsity 50% (Magnitude) COCO2017 bbox: 36.93 (0.40)
segm: 33.23 (0.33)		 Link Link