diff --git a/README.md b/README.md index 6fe14920..a3821cef 100644 --- a/README.md +++ b/README.md @@ -135,7 +135,7 @@ The code is developed under the following configurations. chmod +x demo_test.sh ./demo_test.sh ``` -This script downloads trained models and a test image, runs the test script, and saves predicted segmentation (.png) to the working directory. +This script downloads trained a model (ResNet50dilated + PPM_deepsup) and a test image, runs the test script, and saves predicted segmentation (.png) to the working directory. 2. To test on multiple images, you can simply do something as the following (```$PATH_IMG1, $PATH_IMG2, $PATH_IMG3```are your image paths): ``` @@ -159,7 +159,14 @@ chmod +x download_ADE20K.sh python3 train.py --num_gpus NUM_GPUS ``` -Train a UPerNet101 +* Train ResNet18dilated + PPM_deepsup +```bash +python3 train.py \ + --num_gpus NUM_GPUS arch_encoder resnet18dilated --arch_decoder ppm_deepsup \ + --fc_dim 512 +``` + +* Train UPerNet101 ```bash python3 train.py \ --num_gpus NUM_GPUS --arch_encoder resnet101 --arch_decoder upernet \ @@ -170,14 +177,22 @@ python3 train.py \ ## Evaluation -1. Evaluate a trained network on the validation set. Add ```--visualize``` option to output visualizations as shown in teaser. +1. Evaluate a trained network on the validation set. ```--id``` is the folder name under ```ckpt``` directory. ```--suffix``` defines which checkpoint to use, for example ```_epoch_20.pth```. Add ```--visualize``` option to output visualizations as shown in teaser. ```bash python3 eval.py --id MODEL_ID --suffix SUFFIX ``` -Evaluate a UPerNet (e.g, UPerNet50) + +* Evaluate ResNet18dilated + PPM_deepsup +```bash +python3 eval.py \ + --id MODEL_ID --suffix SUFFIX arch_encoder resnet18dilated --arch_decoder ppm_deepsup \ + --fc_dim 512 +``` + +* Evaluate UPerNet101 ```bash python3 eval.py \ - --id MODEL_ID --suffix SUFFIX --arch_encoder resnet50 --arch_decoder upernet \ + --id MODEL_ID --suffix SUFFIX --arch_encoder resnet101 --arch_decoder upernet \ --padding_constant 32 ```