mirzaevinom / data_science_bowl_2018 Goto Github PK
View Code? Open in Web Editor NEWMy 5th place (out of 816 teams) solution to The 2018 Data Science Bowl organized by Booz Allen Hamilton
My 5th place (out of 816 teams) solution to The 2018 Data Science Bowl organized by Booz Allen Hamilton
Thanks for your interesting repo, it inspired me a lot in my projects.
I think there is a small issue in cluster_prediction function:
data_science_bowl_2018/codes/predict.py
Line 232 in 0a6d0ed
if overlaps[mm].size != 0 and np.max(overlaps[mm]) > 0.1:
Thanks for sharing your code and the procedure with such detail Inom - I have learned a lot from it.
I was wondering if you have made your final trained weights publicly available? I am doing something similar to what you have done, but my training data is a bit small and that's why my results are not great (though they aren't terrible either).
My guess is that if I initiate my training using the weights you have obtained from this competition, I can get better results.
I don't know whether it is common to share the trained weights. If for any reason you cannot do it, I understand.
Thank you!
Excellent repo. I liked the Interesting findings and the Unsuccessful approaches tried. Very useful for future experiments.
I still have one question. Why you didn't train the heads?
Many people recommended to train heads first, then a few times the 4+ (optional) and then train all. Why did your train all since the beginning.
Is there any reason you skipped the train of heads part?
Thank you for making your code public and let me learn.
but i download your code and follow your Instruction,i met some issues
when i run
python augment_preprocess.py
i met some issues:
Traceback (most recent call last):
File "augment_preprocess.py", line 473, in
make_n_save_mosaic()
File "augment_preprocess.py", line 361, in make_n_save_mosaic
with h5py.File(path, "r") as hf:
File "/home/lgf/anaconda3/envs/tensorflow-gpu/lib/python3.6/site-packages/h5py/_hl/files.py", line 312, in init
fid = make_fid(name, mode, userblock_size, fapl, swmr=swmr)
File "/home/lgf/anaconda3/envs/tensorflow-gpu/lib/python3.6/site-packages/h5py/_hl/files.py", line 142, in make_fid
fid = h5f.open(name, flags, fapl=fapl)
File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
File "h5py/h5f.pyx", line 78, in h5py.h5f.open
OSError: Unable to open file (unable to open file: name = '../data/stage1_train/ad473063dab4bf4f2461d9a99a9c0166d4871f156516d9e0a523484e7cf2258d/masks/ad473063dab4bf4f2461d9a99a9c0166d4871f156516d9e0a523484e7cf2258d.h5', errno = 2, error message = 'No such file or directory', flags = 0, o_flags = 0)
i want know what is this .h5 file
And i want to skip preprocess, and run
python train.py
i met some issues again:
Traceback (most recent call last):
File "train.py", line 128, in
train_path, seed=11, test_size=0.1)
File "train.py", line 100, in train_validation_split
df = pd.read_csv('../data/classes.csv')
File "/home/lgf/anaconda3/envs/tensorflow-gpu/lib/python3.6/site-packages/pandas/io/parsers.py", line 678, in parser_f
return _read(filepath_or_buffer, kwds)
File "/home/lgf/anaconda3/envs/tensorflow-gpu/lib/python3.6/site-packages/pandas/io/parsers.py", line 440, in _read
parser = TextFileReader(filepath_or_buffer, **kwds)
File "/home/lgf/anaconda3/envs/tensorflow-gpu/lib/python3.6/site-packages/pandas/io/parsers.py", line 787, in init
self._make_engine(self.engine)
File "/home/lgf/anaconda3/envs/tensorflow-gpu/lib/python3.6/site-packages/pandas/io/parsers.py", line 1014, in _make_engine
self._engine = CParserWrapper(self.f, **self.options)
File "/home/lgf/anaconda3/envs/tensorflow-gpu/lib/python3.6/site-packages/pandas/io/parsers.py", line 1708, in init
self._reader = parsers.TextReader(src, **kwds)
File "pandas/_libs/parsers.pyx", line 384, in pandas._libs.parsers.TextReader.cinit
File "pandas/_libs/parsers.pyx", line 695, in pandas._libs.parsers.TextReader._setup_parser_source
FileNotFoundError: File b'../data/classes.csv' does not exist
I can't run again because I don't have a .csv.
Hi @mirzaevinom,
could you please briefly describe the modifications that you applied on matterports implementation?
Maybe you could write it somewhere in the documentation
kkkk
When I use mask rcnn, I get a bad performance on masks. The mask doesn't match well with the edge of bbox. The predicted bbox coverd the object in the image, but the pred_mask can,t cover the object. What should I do? To optimize the code of mask generator? The image in dataset has a size of 1028*1280 with a large object. Thank you very much
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.