@@ -8,11 +8,10 @@ Features:
88- Distributed training with native automatic mixed precision
99- Experiments tracking with [ ClearML] ( https://github.com/allegroai/clearml )
1010
11- Experiment | Model | Dataset | Val Avg IoU | ClearML Link
12- ---|---|---|---|---
13- configs/baseline_dplv3_resnet101.py | DeepLabV3 Resnet101 | VOC Only | 0.659161 | [ link] ( https://app.clear.ml/projects/0e9a3a92d3134283b7d5572d516d60c5/experiments/a7254f084a9e47ca9380dfd739f89520/output/execution )
14- configs/baseline_dplv3_resnet101_sbd.py | DeepLabV3 Resnet101 | VOC+SBD | 0.6853087 | [ link] ( https://app.clear.ml/projects/0e9a3a92d3134283b7d5572d516d60c5/experiments/dc4cee3377a74d19bc2d0e0e4d638c1f/output/execution )
15-
11+ | Experiment | Model | Dataset | Val Avg IoU | ClearML Link |
12+ | --------------------------------------- | ------------------- | -------- | ----------- | ------------------------------------------------------------------------------------------------------------------------------------ |
13+ | configs/baseline_dplv3_resnet101.py | DeepLabV3 Resnet101 | VOC Only | 0.659161 | [ link] ( https://app.clear.ml/projects/0e9a3a92d3134283b7d5572d516d60c5/experiments/a7254f084a9e47ca9380dfd739f89520/output/execution ) |
14+ | configs/baseline_dplv3_resnet101_sbd.py | DeepLabV3 Resnet101 | VOC+SBD | 0.6853087 | [ link] ( https://app.clear.ml/projects/0e9a3a92d3134283b7d5572d516d60c5/experiments/dc4cee3377a74d19bc2d0e0e4d638c1f/output/execution ) |
1615
1716## Setup
1817
@@ -23,10 +22,13 @@ pip install -r requirements.txt
2322### Docker
2423
2524For docker users, you can use the following images to run the example:
25+
2626``` bash
2727docker pull pytorchignite/vision:latest
2828```
29+
2930or
31+
3032``` bash
3133docker pull pytorchignite/hvd-vision:latest
3234```
@@ -51,7 +53,6 @@ This script will download and extract the following datasets into `/path/to/data
5153- The [ Pascal VOC2012] ( http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar ) dataset
5254- Optionally, the [ SBD] ( http://www.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/semantic_contours/benchmark.tgz ) evaluation dataset
5355
54-
5556## Usage
5657
5758Please, export the ` DATASET_PATH ` environment variable for the Pascal VOC2012 dataset.
@@ -75,6 +76,7 @@ export SBD_DATASET_PATH=/path/to/SBD/
7576- Adjust batch size for your GPU type in the configuration file: ` configs/baseline_dplv3_resnet101_sbd.py ` or ` configs/baseline_dplv3_resnet101.py `
7677
7778Run the following command:
79+
7880``` bash
7981CUDA_VISIBLE_DEVICES=0 python -u main.py training configs/baseline_dplv3_resnet101_sbd.py
8082# or without SBD
@@ -121,7 +123,6 @@ torchrun --nproc_per_node=2 main.py eval configs/eval_baseline_dplv3_resnet101_s
121123horovodrun -np=2 python -u main.py eval configs/eval_baseline_dplv3_resnet101_sbd.py --backend=" horovod"
122124```
123125
124-
125126## Acknowledgements
126127
127128Trainings were done using credits provided by AWS for open-source development via NumFOCUS
0 commit comments