Added script files for setup the project on Windows OS

README.md

@@ -1,3 +1,5 @@
+Linux | [Windows](https://github.com/zumrudu-anka/SiamMask/blob/master/READMEForWindows.md)
+
 # SiamMask
 
 **NEW:** now including code for both training and inference!
@@ -29,12 +31,11 @@ If you find this code useful, please consider citing:
 }
 ```
 
-
 ## Contents
 1. [Environment Setup](#environment-setup)
 2. [Demo](#demo)
-3. [Testing Models](#testing-models)
-4. [Training Models](#training-models)
+3. [Testing](#testing)
+4. [Training](#training)
 
 ## Environment setup
 This code has been tested on Ubuntu 16.04, Python 3.6, Pytorch 0.4.1, CUDA 9.2, RTX 2080 GPUs

READMEForWindows.md

@@ -0,0 +1,139 @@
+[Linux](https://github.com/zumrudu-anka/SiamMask/blob/master/README.md) | Windows
+
+## Contents
+1. [Environment Setup](#environment-setup)
+2. [Demo](#demo)
+3. [Testing](#testing)
+4. [TODO LIST](#todo-list)
+5. [References](#references)
+
+## Environment Setup
+- Clone the [Repository](https://github.com/foolwood/SiamMask)
+```
+git clone https://github.com/foolwood/SiamMask.git
+set SiamMask=%cd%
+```
+- Setup Python Environment
+```
+python -m venv venv
+venv\Scripts\activate
+pip install -r requirements.txt 
+python make.py
+```
+> If you see any error when install packages. Check it and please install up-to-date versions of the packages
+- Add the project to your PYTHONPATH
+```
+set PYTHONPATH=%PYTHONPATH%;<your project path>
+```
+## Demo
+- [Setup](#environment-setup) your environment
+- Download the SiamMask model
+```
+cd <your project directory>/experiments/siammask_sharp
+python downloadSiamMaskModel.py
+```
+- Change `tools\demo.py` 26. line `from custom import Custom` to `from experiments.siammask_sharp.custom import Custom`
+- Change `experiments\siammask_sharp\custom.py` 9. line `from resnet import resnet50` to `from .resnet import resnet50`
+- Run demo.py
+```
+cd <your project directory>/experiments/siammask_sharp
+set PYTHONPATH=%PYTHONPATH%;<your project path>
+python ../../tools/demo.py --resume SiamMask_DAVIS.pth --config config_davis.json
+```
+## Testing
+- [Setup](#environment-setup) your environment
+- Download test data
+```
+cd <your project directory>/data
+python get_test_data.py
+```
+- Download pretrained models
+```
+cd <your project directory>/experiments/siammask_sharp
+python downloadSiamMaskModel.py
+python ../../tools/demo.py --resume SiamMask_DAVIS.pth --config config_davis.json --base_path ../../data/VOT2019/drone_across
+```
+
+## TODO LIST
+
+- Evaluate performance on [VOT](http://www.votchallenge.net/)
+```
+....
+```
+
+- Evaluate performance on [DAVIS](https://davischallenge.org/) (less than 50s)
+```
+....
+```
+- Evaluate performance on [Youtube-VOS](https://youtube-vos.org/) (need download data from [website](https://youtube-vos.org/dataset/download))
+```
+....
+```
+
+- ## Training
+
+- ### Training Data 
+ - Download the [Youtube-VOS](https://youtube-vos.org/dataset/download/), 
+[COCO](http://cocodataset.org/#download), 
+[ImageNet-DET](http://image-net.org/challenges/LSVRC/2015/), 
+and [ImageNet-VID](http://image-net.org/challenges/LSVRC/2015/).
+ - Preprocess each datasets according the [readme](data/coco/readme.md) files.
+
+- ### Download the pre-trained model (174 MB)
+ (This model was trained on the ImageNet-1k Dataset)
+ ```
+ cd $SiamMask/experiments
+ wget http://www.robots.ox.ac.uk/~qwang/resnet.model
+ ls | grep siam | xargs -I {} cp resnet.model {}
+ ```
+
+- ### Training SiamMask base model
+ - [Setup](#environment-setup) your environment
+ - From the experiment directory, run
+ ```
+ cd $SiamMask/experiments/siammask_base/
+ bash run.sh
+ ```
+ - Training takes about 10 hours in our 4 Tesla V100 GPUs.
+ - If you experience out-of-memory errors, you can reduce the batch size in `run.sh`.
+ - You can view progress on Tensorboard (logs are at <experiment\_dir>/logs/)
+ - After training, you can test checkpoints on VOT dataset.
+ ```shell
+ bash test_all.sh -s 1 -e 20 -d VOT2018 -g 4 # test all snapshots with 4 GPUs
+ ```
+ - Select best model for hyperparametric search.
+ ```shell
+ #bash test_all.sh -m [best_test_model] -d VOT2018 -n [thread_num] -g [gpu_num] # 8 threads with 4 GPUS
+ bash test_all.sh -m snapshot/checkpoint_e12.pth -d VOT2018 -n 8 -g 4 # 8 threads with 4 GPUS
+ ```
+
+- ### Training SiamMask model with the Refine module
+ - [Setup](#environment-setup) your environment
+ - In the experiment file, train with the best SiamMask base model
+ ```
+ cd $SiamMask/experiments/siammask_sharp
+ bash run.sh <best_base_model>
+ bash run.sh checkpoint_e12.pth
+ ```
+ - You can view progress on Tensorboard (logs are at <experiment\_dir>/logs/)
+ - After training, you can test checkpoints on VOT dataset
+ ```shell
+ bash test_all.sh -s 1 -e 20 -d VOT2018 -g 4
+ ```
+
+- ### Training SiamRPN++ model (*unofficial*)
+ - [Setup](#environment-setup) your environment
+ - From the experiment directory, run
+ ```
+ cd $SiamMask/experiments/siamrpn_resnet
+ bash run.sh
+ ```
+ - You can view progress on Tensorboard (logs are at <experiment\_dir>/logs/)
+ - After training, you can test checkpoints on VOT dataset
+ ```shell
+ bash test_all.sh -h
+ bash test_all.sh -s 1 -e 20 -d VOT2018 -g 4
+ ```
+
+### References
+[SiamMask](https://github.com/foolwood/SiamMask)

data/filesForMoveToTrackdat/download_vot.py

@@ -0,0 +1,75 @@
+import sys
+import os
+import requests
+import json
+import zipfile
+import io
+import urllib.request, shutil
+
+VOT_YEAR = sys.argv[1]
+VOT_CHALLENGE = "main"
+name = "vot" + VOT_YEAR + "_" + VOT_CHALLENGE
+
+dl = f"../dl/{name}"
+
+try:
+ os.makedirs(dl)
+except FileExistsError:
+ print("Loading...")
+
+abspath = os.path.abspath(__file__)
+projectDir = os.path.dirname(abspath)
+
+os.chdir(dl)
+
+base_url = f"https://data.votchallenge.net/vot{VOT_YEAR}/{VOT_CHALLENGE}"
+
+if os.path.exists("description.json"):
+ pass
+else:
+ response = requests.get(base_url + "/description.json", allow_redirects = True)
+ with open("description.json", "wb") as descriptionFile:
+ descriptionFile.write(response.content)
+
+with open("annotations.txt", "w") as annotationsFileOutput, open("color.txt", "w") as colorFileOutput, open("list.txt", "w") as listFileOutput:
+ with open("description.json", "r") as description:
+ data = json.load(description)
+ for sequence in data["sequences"]:
+ annotationsFileOutput.write(sequence["annotations"]["url"] + "\n")
+ colorFileOutput.write(sequence["channels"]["color"]["url"] + "\n")
+ listFileOutput.write(sequence["name"] + "\n")
+
+try:
+ os.makedirs("annotations")
+except FileExistsError:
+ print("Loading...")
+
+try:
+ os.makedirs("color")
+except FileExistsError:
+ print("Loading...")
+
+os.chdir("./annotations")
+with open("../annotations.txt", "r") as annotationsTxtFile:
+ lines = annotationsTxtFile.readlines()
+ for line in lines:
+ url = f"{base_url}/{line}"
+ response = requests.get(url, stream = True)
+ with urllib.request.urlopen(url) as response, open(line.strip(), "wb") as downloadedAnnotation:
+ shutil.copyfileobj(response, downloadedAnnotation)
+ print("Test datas are downloading...")
+
+os.chdir("../color")
+with open("../color.txt", "r") as colorTxtFile:
+ lines = colorTxtFile.readlines()
+ for line in lines:
+ url = f"{base_url}/{line}"
+ response = requests.get(url, stream = True)
+ with urllib.request.urlopen(url) as response, open(line[16:].strip(), "wb") as downloadedColor:
+ shutil.copyfileobj(response, downloadedColor)
+ print("Test datas are downloading...")
+
+ # response = requests.get(url, allow_redirects = True)
+ # with open(line[16:].strip(), "wb") as downloadedColor:
+ # downloadedColor.write(response.content)
+ # print("Test datas are downloading...")

data/filesForMoveToTrackdat/unpack_vot.py

@@ -0,0 +1,22 @@
+import sys
+import os
+from shutil import copyfile
+VOT_YEAR = sys.argv[1]
+VOT_CHALLENGE = "main"
+
+defaultName = f"vot{VOT_YEAR}_{VOT_CHALLENGE}"
+
+dl = "../" + sys.argv[2]
+data = "../../" + sys.argv[3]
+# scripts="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
+
+try:
+ os.makedirs(data)
+except FileExistsError:
+ print("Loading...")
+
+descriptionFileSrc = f"../dl/{defaultName}/description.json"
+descriptionFileDestination = f"{data}/description.json"
+copyfile(descriptionFileSrc, descriptionFileDestination)
+
+os.system(f"python unzip_vot.py {dl} {data}")

data/get_test_data.py

@@ -0,0 +1,65 @@
+import os
+from shutil import copyfile
+import requests
+
+abspath = os.path.abspath(__file__)
+projectDir = os.path.dirname(abspath)
+trackDat = projectDir + "/trackdat"
+
+if not os.path.exists(trackDat):
+ os.system("git clone https://github.com/jvlmdr/trackdat.git")
+else:
+ print("Loading...")
+
+scripts = trackDat + "/scripts"
+
+copyfile(projectDir + "/filesForMoveToTrackdat/download_vot.py", scripts + "/download_vot.py")
+copyfile(projectDir + "/filesForMoveToTrackdat/unpack_vot.py", scripts + "/unpack_vot.py")
+
+os.chdir(scripts)
+os.system("download_vot.py 2016 dl/vot2016")
+os.system("download_vot.py 2018 dl/vot2018")
+os.system("download_vot.py 2019 dl/vot2019")
+os.system("unpack_vot.py 2016 dl/vot2016_main VOT2016")
+os.system("unpack_vot.py 2018 dl/vot2018_main VOT2018")
+os.system("unpack_vot.py 2019 dl/vot2019_main VOT2019")
+
+Vot2016ListTxtPath = trackDat + "/dl/vot2016_main/list.txt"
+Vot2018ListTxtPath = trackDat + "/dl/vot2018_main/list.txt"
+Vot2019ListTxtPath = trackDat + "/dl/vot2019_main/list.txt"
+
+copyfile(Vot2016ListTxtPath, projectDir + "/VOT2016/list.txt")
+copyfile(Vot2018ListTxtPath, projectDir + "/VOT2018/list.txt")
+copyfile(Vot2019ListTxtPath, projectDir + "/VOT2019/list.txt")
+
+os.chdir(projectDir)
+
+base_url = "http://www.robots.ox.ac.uk/~qwang/"
+urlArray = [
+ "VOT2016.json",
+ "VOT2018.json",
+]
+
+for url in urlArray:
+ response = requests.get(base_url + url, allow_redirects = True)
+ with open(url, "wb") as jsonFile:
+ jsonFile.write(response.content)
+
+
+
+
+# ! TODOOOOO
+
+# # json file for eval toolkit
+# wget http://www.robots.ox.ac.uk/~qwang/VOT2016.json
+# wget http://www.robots.ox.ac.uk/~qwang/VOT2018.json
+# python create_json.py VOT2019
+
+# # DAVIS
+# wget https://data.vision.ee.ethz.ch/csergi/share/davis/DAVIS-2017-trainval-480p.zip
+# unzip DAVIS-2017-trainval-480p.zip
+# ln -s ./DAVIS ./DAVIS2016
+# ln -s ./DAVIS ./DAVIS2017
+
+
+# # Youtube-VOS

experiments/siammask_sharp/downloadSiamMaskModel.py

@@ -0,0 +1,28 @@
+import requests
+import os
+#? Basic Example
+
+# url = 'https://www.facebook.com/favicon.ico'
+# r = requests.get(url, allow_redirects=True)
+
+# open('facebook.ico', 'wb').write(r.content)
+
+#? Basic Example
+
+#* Running Part
+
+base_url = 'http://www.robots.ox.ac.uk/~qwang/'
+
+urlArray = [
+ "SiamMask_VOT.pth",
+ "SiamMask_VOT_LD.pth",
+ "SiamMask_DAVIS.pth"
+]
+
+for url in urlArray:
+ print("Pretrained models are downloading...")
+ if not os.path.exists(url):
+ r = requests.get(base_url + url, allow_redirects = True)
+ open(url, 'wb').write(r.content)
+
+print("Download complete...")

make.py

@@ -0,0 +1,15 @@
+import os
+
+abspath = os.path.abspath(__file__)
+projectDir = os.path.dirname(abspath)
+
+pyVotkitDir = "/utils/pyvotkit"
+pySotDir = "/utils/pysot/utils"
+
+os.chdir(projectDir + pyVotkitDir)
+os.system("python setup.py build_ext --inplace")
+
+os.chdir(projectDir + pySotDir)
+os.system("python setup.py build_ext --inplace")
+
+os.chdir(projectDir)

requirements.txt

@@ -3,7 +3,6 @@ colorama==0.3.9
 numpy==1.15.4
 requests==2.21.0
 fire==0.1.3
-torch==0.4.1
 matplotlib==2.2.3
 numba==0.39.0
 scipy==1.1.0
@@ -12,5 +11,5 @@ pandas==0.23.4
 tqdm==4.29.1
 tensorboardX==1.6
 opencv_python==3.4.3.18
-torch==0.4.1
-torchvision==0.2.1
+
+# pip install torch===1.4.0 torchvision===0.5.0 -f https://download.pytorch.org/whl/torch_stable.html