Performance focused implementation of Mask RCNN based on the Tensorpack implementation. The original paper: Mask R-CNN
This implementation of Mask RCNN is focused on increasing training throughput without sacrificing any accuracy. We do this by training with a batch size > 1 per GPU using FP16 and two custom TF ops.
Training on N GPUs (V100s in our experiments) with a per-gpu batch size of M = NxM training
Training converges to target accuracy for configurations from 8x1 up to 32x4 training. Training throughput is substantially improved from original Tensorpack code.
A pre-built dockerfile is available in DockerHub under fewu/mask-rcnn-tensorflow:master-latest
. It is automatically built on each commit to master.
- Running this codebase requires a custom TF binary - available under GitHub releases (custom ops and fix for bug introduced in TF 1.13
- We give some details the codebase and optimizations in
CODEBASE.md
- Data preprocessing
- We are using COCO 2017, you can download the data from COCO data.
- The pre-trained resnet backbone can be donloaded from ImageNet-R50-AlignPadding.npz
- The file folder needs to have the following directory structure:
data/ annotations/ instances_train2017.json instances_val2017.json pretrained-models/ ImageNet-R50-AlignPadding.npz train2017/ # image files that are mentioned in the corresponding json val2017/ # image files that are mentioned in corresponding json
- If you want to use COCO 2014, please refer to here
- If you want to use EKS or Sagemaker, you need to create your own S3 bucket which contains the data in the same directory structure, and change the S3 bucket name in the following files:
- EKS: stage-data
- SageMaker: S3 download
- If you want to use EKS, you also need to create the a FSx filesystem
- You don't need to link your S3 bucket if you have followed the previous steps
- You need to change the FSx filesystem id in pv-fsx file.
- Container is recommended for training
The result was running on P3dn.24xl instances using EKS. 12 epochs training:
Num_GPUs x Images_Per_GPU | Training time | Box mAP | Mask mAP |
---|---|---|---|
8x4 | 5.09h | 37.47% | 34.45% |
16x4 | 3.11h | 37.41% | 34.47% |
32x4 | 1.94h | 37.20% | 34.25% |
24 epochs training:
Num_GPUs x Images_Per_GPU | Training time | Box mAP | Mask mAP |
---|---|---|---|
8x4 | 9.78h | 38.25% | 35.08% |
16x4 | 5.60h | 38.44% | 35.18% |
32x4 | 3.33h | 38.33% | 35.12% |
Forked from the excellent Tensorpack repo at commit a9dce5b220dca34b15122a9329ba9ff055e8edc6