def main():
# The basic core takes as input frames from the two camera viewpoints that look into the playing field
# and learns proper kernels for recognizing the scene
# This is a baseline to use as a pretrained classifier for recognizing entire playtypes in the advanced core
global args, mylogger
args = parser.parse_args()
mylogger = MagnificentOracle()
mylogger.set_log(logfile=args.log)
mylogger.log("-dotted-line")
EB = True
postprocess = False
IoU = False
if not EB:
if ("postprocess" not in args.log):
if ("IoU" not in args.log):
vp = viewpoint_classifier_ED(args.arch, path=args.dataset_path, viewpoints = args.viewpoints)
train(int(args.epochs), vp)
else:
print("IoU scores per class")
IoU()
else:
print("post-processing")
post_process()
else:
vp = viewpoint_classifier_ED(args.arch, path=args.dataset_path, viewpoints = args.viewpoints)
EB_extraction(vp)
def main():
global args, mylogger
args = parser.parse_args()
mylogger = MagnificentOracle()
mylogger.set_log(logfile=args.log)
mylogger.log("-dotted-line")
pt = playtype_classifier()
train(int(args.epochs), pt)
def main():
global args, mylogger
args = parser.parse_args()
mylogger = MagnificentOracle()
mylogger.set_log(logfile=args.log)
#mylogger.set_log(logfile=None)
mylogger.log("-dotted-line")
pt = playtype_classifier(dataset_index=int(args.dataset_index))
#train(int(args.epochs), pt)
#final_expectation(pt)
#acc_test(1,pt) # 91% testing acc at 48 epochs for binary, 85% testing acc at 48 epochs for trinary
final_expectation_dataset(pt)
def main():
global mylogger
mylogger = MagnificentOracle()
mylogger.set_log(logfile=None)
mylogger.log("-dotted-line")
kmeans_init = True
word2vec_init = True
if word2vec_init:
word2vec()
elif kmeans_init:
cluster_raw()
else:
c = clustering()
c.run()
def main():
global args, mylogger
args = parser.parse_args()
mylogger = MagnificentOracle()
mylogger.set_log(logfile=args.log)
mylogger.log("-dotted-line")
customset_test = CustomMasterDatasetPlays(path=args.dataset_path,
subset_type="testing",
dataset_index=0,
categories="chunked")
test_acc_loader = torch.utils.data.DataLoader(dataset=customset_test,
batch_size=16,
shuffle=False,
num_workers=args.num_workers)
print len(customset_test), len(test_acc_loader)
def main():
global args, mylogger
args = parser.parse_args()
mylogger = MagnificentOracle()
mylogger.set_log(logfile=args.log)
mylogger.log("-dotted-line")
#ground_truth_files()
pt = playtype_classifier()
#print("All done")
#IoU()
#train(int(args.epochs), pt)
#final_expectation(pt)
#acc_test(1,pt) # 91% testing acc at 48 epochs for binary, 85% testing acc at 48 epochs for trinary
final_expectation_dataset(pt)
def main():
# The basic core takes as input frames from the two camera viewpoints that look into the playing field
# and learns proper kernels for recognizing the scene
# This is a baseline to use as a pretrained classifier for recognizing entire playtypes in the advanced core
global args, mylogger
args = parser.parse_args()
mylogger = MagnificentOracle()
mylogger.set_log(logfile=args.log)
mylogger.log("-dotted-line")
# TO-DO
vp = viewpoint_classifier_ED(args.arch,
path=args.dataset_path,
viewpoints=args.viewpoints)
train(int(args.epochs), vp)
def main():
# The basic core takes as input frames from the two camera viewpoints that look into the playing field
# and learns proper kernels for recognizing the scene
# This is a baseline to use as a pretrained classifier for recognizing entire playtypes in the advanced core
global args, mylogger
args = parser.parse_args()
mylogger = MagnificentOracle()
mylogger.set_log(logfile=args.log)
mylogger.log("-dotted-line")
EB = False
postprocess = False
IoU_ = True
if not EB:
if not postprocess:
if not IoU_:
vp = viewpoint_classifier_ED(args.arch,
path=args.dataset_path,
viewpoints=args.viewpoints)
train(int(args.epochs), vp)
else:
for folder in [
"ED32_vgg_frozen"
]: #["framewise","ED32","ED196","ED32_vgg_frozen"]:
IoU(folder=folder)
else:
for folder in ["ED32_vgg_frozen"
]: #["framewise","ED32","ED196","ED32_vgg_frozen"]:
print("post-processing")
post_process(folder=folder)
else:
vp = viewpoint_classifier_ED(args.arch,
path=args.dataset_path,
viewpoints=args.viewpoints)
EB_extraction(vp)