def index():
upload = request.files.get("files")
path = "uploads/" + hashlib.sha256(
os.urandom(16)).hexdigest()[:16] + ".png"
upload.save(path)
print("yes")
return prediction.run()
def run(argv=None):
if argv is None or argv[1] == 'prediction':
prediction.run(argv[1:])
elif argv[1] == 'select':
selection.run(argv[1:])
elif argv[1] == 'extract':
extraction.run(argv[1:])
elif argv[1] == 'extract_UKBB_v1':
extraction.run_UKBB_v1(argv[1:])
elif argv[1] == 'extract_UKBB_v2':
extraction.run_UKBB_v2(argv[1:])
elif argv[1] == 'test_fit_time':
test_fit_time.run(argv[1:])
elif argv[1] == 'stats':
statistics.run(argv[1:])
else:
raise ValueError(f'Unrecognized argument {argv[1]}')
def run_one(task, T):
argv = {
'action': 'prediction',
'task_name': task,
'strategy_name': '0',
'T': str(T),
'RS': '0',
'dump_idx_only': True,
'n_top_pvals': 100,
}
# Only one trial for task having features manually selected (not _pvals)
if '_pvals' not in task and T != 0:
return
args = Namespace(**argv)
prediction.run(args)
def build_model(tech_list):
run(tech_list)
def test_run(self):
print "Test prediction required run TestServerMonitor.py and TestUserMonitor.py first"
pr.run("MONITOR_TEST")
def stats(model, dataset):
print(performance(model))
for i in range(2100):
img, p, pose = dataset.__getitem__(i)
run(model, img, i)
import torch
import os
from prediction import load, run, performance
def stats(model, dataset):
print(performance(model))
for i in range(2100):
img, p, pose = dataset.__getitem__(i)
run(model, img, i)
# model, dataset = load("boxClassifier")
# stats(model, dataset)
model, dataset = load("boxClassifier")
for i in range(20):
img, p, pose = dataset.__getitem__(i)
run(model, img, i)