Uploaded create_sample_submission.py
Browse files- create_sample_submission.py +66 -0
create_sample_submission.py
ADDED
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import numpy as np
|
2 |
+
import pandas as pd
|
3 |
+
import h5py
|
4 |
+
|
5 |
+
from trimesh.voxel.runlength import dense_to_brle
|
6 |
+
from pathlib import Path
|
7 |
+
from collections import defaultdict
|
8 |
+
|
9 |
+
from typing import Any, Union, Dict, Literal
|
10 |
+
from numpy.typing import NDArray
|
11 |
+
|
12 |
+
# class RandomModel:
|
13 |
+
# def __init__(self, shape):
|
14 |
+
# self.shape = shape
|
15 |
+
# return
|
16 |
+
|
17 |
+
# def __call__(self, input):
|
18 |
+
# # input is ignored, just generate some random predictions
|
19 |
+
# return np.random.randint(0, 2, size=self.shape, dtype=bool)
|
20 |
+
|
21 |
+
class FixedModel:
|
22 |
+
def __init__(self, shape) -> None:
|
23 |
+
self.shape = shape
|
24 |
+
return
|
25 |
+
|
26 |
+
def __call__(self, input) -> Any:
|
27 |
+
# input is ignored, just generate a mask filled with zeros, with fixed pixels set to 1
|
28 |
+
mask = np.zeros(self.shape, dtype=bool)
|
29 |
+
mask[100:250, 100:250] = True
|
30 |
+
return mask
|
31 |
+
|
32 |
+
def retrieve_validation_fold(path: Union[str, Path]) -> Dict[str, NDArray]:
|
33 |
+
result = defaultdict(dict)
|
34 |
+
with h5py.File(path, 'r') as fp:
|
35 |
+
for uuid, values in fp.items():
|
36 |
+
if values.attrs['fold'] != 0:
|
37 |
+
continue
|
38 |
+
|
39 |
+
result[uuid]['post'] = values['post_fire'][...]
|
40 |
+
# result[uuid]['pre'] = values['pre_fire'][...]
|
41 |
+
|
42 |
+
return dict(result)
|
43 |
+
|
44 |
+
def compute_submission_mask(id: str, mask: NDArray):
|
45 |
+
brle = dense_to_brle(mask.astype(bool).flatten())
|
46 |
+
return {"id": id, "rle_mask": brle, "index": np.arange(len(brle))}
|
47 |
+
|
48 |
+
if __name__ == '__main__':
|
49 |
+
validation_fold = retrieve_validation_fold('train_eval.hdf5')
|
50 |
+
|
51 |
+
# use a list to accumulate results
|
52 |
+
result = []
|
53 |
+
# instantiate the model
|
54 |
+
model = FixedModel(shape=(512, 512))
|
55 |
+
for uuid in validation_fold:
|
56 |
+
input_images = validation_fold[uuid]
|
57 |
+
|
58 |
+
# perform the prediction
|
59 |
+
predicted = model(input_images)
|
60 |
+
# convert the prediction in RLE format
|
61 |
+
encoded_prediction = compute_submission_mask(uuid, predicted)
|
62 |
+
result.append(pd.DataFrame(encoded_prediction))
|
63 |
+
|
64 |
+
# concatenate all dataframes
|
65 |
+
submission_df = pd.concat(result)
|
66 |
+
submission_df.to_csv('predictions.csv', index=False)
|