mm/citybert
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Compare commits

base: mm:294d4bb1cd68bb6f91f7d2a7c0f4b8aca18655ad
Branches Tags
mm:main
..
compare: mm:e924ba422746b0e7cc00ae4e73a4bbaf28d58c49
Branches Tags
mm:main

1 Commits
294d4bb1cd ... e924ba4227

Author SHA1 Message Date
mm
e924ba4227 progress bar 2023-05-05 06:29:56 +00:00
2 changed files with 1 additions and 8 deletions
Show Stats Expand all files Collapse all files

2
Makefile
View File

@ -2,7 +2,7 @@ all: install data train eval
city_distances_full.csv: check generate_data.py city_distances_full.csv: check generate_data.py
@echo "Generating distance data..." @echo "Generating distance data..."
@bash -c 'time python generate_data.py --country US --workers 8 --chunk-size 4200' @bash -c 'time python generate_data.py --country US --workers 8 --chunk-size 8000'
data: city_distances_full.csv data: city_distances_full.csv

7
generate_data.py
View File

@ -34,11 +34,6 @@ parser.add_argument(
type=str, type=str,
default="city_distances_full.csv", default="city_distances_full.csv",
) )
parser.add_argument(
"--shuffle",
action="store_true",
help="Option to shuffle combinations list before iterating over it",
)
args = parser.parse_args() args = parser.parse_args()
@ -121,8 +116,6 @@ def main():
cities = list(us_cities.values()) cities = list(us_cities.values())
print(f"Num cities: {len(cities)}") print(f"Num cities: {len(cities)}")
city_combinations = list(itertools.combinations(cities, 2)) city_combinations = list(itertools.combinations(cities, 2))
if args.shuffle:
np.random.shuffle(city_combinations)
chunk_size = args.chunk_size chunk_size = args.chunk_size
num_chunks = len(city_combinations) // chunk_size + 1 num_chunks = len(city_combinations) // chunk_size + 1
output_file = args.output_file output_file = args.output_file