-
Notifications
You must be signed in to change notification settings - Fork 3
/
make_batches.py
43 lines (36 loc) · 1.52 KB
/
make_batches.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
import gzip
import io
import json
import os
import habitat
def make_batches(scenes, out_dir, num_episodes_per_scene,total_batches):
for i in range(total_batches):
ep_id =0
dset_all = habitat.datasets.make_dataset("PointNav-v1")
for j in range(len(scenes)):
source_dataset_path = scenes[j]
with gzip.open(source_dataset_path, "rt") as f:
deserialized = json.loads(f.read())
# if j =0 thn
dset = habitat.datasets.make_dataset("PointNav-v1")
dset.episodes = list(deserialized['episodes'][num_episodes_per_scene*i: num_episodes_per_scene*i +num_episodes_per_scene])
for k, ep in enumerate(dset.episodes):
ep['episode_id'] = str(ep_id)
ep_id +=1
dset_all.episodes.extend(dset.episodes)
out_file = out_dir+'training_batch_'+str(i)+'.json.gz'
os.makedirs(osp.dirname(out_file), exist_ok=True)
with gzip.open(out_file, "wt") as f:
f.write(dset_all.to_json())
def main():
# Specify scenes directory with the following structure
# -- train/val/test
# ------content
# ----------env_name.json.gz
scenes = glob.glob("/home/mirshad7/habitat-api/data/datasets/pointnav/gibson/v1/train/content/*.json.gz")
out_dir = f'../data/datasets/pointnav/gibson/v1/all_val/'
num_episodes_per_scene = 2
total_batches = 100
make_batches(scenes, out_dir,num_episodes_per_scene,total_batches)
if __name__ == "__main__":
main()