-
Notifications
You must be signed in to change notification settings - Fork 0
/
submit.py
85 lines (64 loc) · 2.82 KB
/
submit.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
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
"""The main script for running inference on the test set and creating a submission."""
import os
import warnings
from pathlib import Path
import hydra
from epochalyst.logging.section_separator import print_section_separator
from hydra.core.config_store import ConfigStore
from omegaconf import DictConfig
from src.config.submit_config import SubmitConfig
from src.setup.setup_data import setup_inference_data
from src.setup.setup_pipeline import setup_pipeline
from src.setup.setup_runtime_args import setup_pred_args
from src.utils.logger import logger
from src.utils.set_torch_seed import set_torch_seed
from src.utils.to_submission_format import to_submission_format
warnings.filterwarnings("ignore", category=UserWarning)
# Makes hydra give full error messages
os.environ["HYDRA_FULL_ERROR"] = "1"
# os.environ["CUDA_VISIBLE_DEVICES"] = "-1"
# Set up the config store, necessary for type checking of config yaml
cs = ConfigStore.instance()
cs.store(name="base_submit", node=SubmitConfig)
@hydra.main(version_base=None, config_path="conf", config_name="submit")
# TODO(Epoch): Use SubmitConfig instead of DictConfig
def run_submit(cfg: DictConfig) -> None:
"""Run the main script for submitting the predictions.
:param cfg: The config object. Created with Hydra.
:raise ValueError: If predictions are None.
"""
print_section_separator("Q4 - BirdCLEF - Submit")
# Set up logging
try:
import coloredlogs
coloredlogs.install()
except ImportError:
"""Coloredlogs is not installed."""
set_torch_seed()
# Get output directory
output_dir = Path(hydra.core.hydra_config.HydraConfig.get().runtime.output_dir)
# Preload the pipeline
print_section_separator("Setup pipeline")
model_pipeline = setup_pipeline(cfg, is_train=False)
# Load the test data
X = setup_inference_data(cfg.data_path)
# Predict on the test data
logger.info("Making predictions...")
pred_args = setup_pred_args(pipeline=model_pipeline, output_dir=output_dir.as_posix(), data_dir=cfg.data_path, species_dir=cfg.species_path)
predictions = model_pipeline.predict(X, **pred_args)
# Make submission
if predictions is not None:
# Create a dataframe from the predictions
submission = to_submission_format(predictions, cfg.data_path, cfg.species_path)
# Print submission head
logger.info(submission.head())
# Save submissions to path (Might be different for other platforms than Kaggle)
result_path = Path(cfg.result_path)
os.makedirs(result_path, exist_ok=True)
submission_path = result_path / "submission.csv"
submission.to_csv(submission_path, index=False)
logger.info(f"Submission saved to {submission_path}")
else:
raise ValueError("Predictions are None")
if __name__ == "__main__":
run_submit()