Bienqq
/
car-detection-bayes
2
1
Fork 0
Code Issues Pull Requests Releases Wiki Activity

updates

This commit is contained in:
Glenn Jocher 2019-02-23 23:50:23 +01:00
parent 9e60e97a6c
commit 8af70386e8
1 changed files with 32 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

37
test.py
View File

@ -37,7 +37,7 @@ def test(
# dataloader = torch.utils.data.DataLoader(LoadImagesAndLabels(test_path), batch_size=batch_size) # pytorch # dataloader = torch.utils.data.DataLoader(LoadImagesAndLabels(test_path), batch_size=batch_size) # pytorch
dataloader = LoadImagesAndLabels(test_path, batch_size=batch_size, img_size=img_size) dataloader = LoadImagesAndLabels(test_path, batch_size=batch_size, img_size=img_size)
mean_mAP, mean_R, mean_P = 0.0, 0.0, 0.0 mean_mAP, mean_R, mean_P, seen = 0.0, 0.0, 0.0, 0
print('%11s' * 5 % ('Image', 'Total', 'P', 'R', 'mAP')) print('%11s' * 5 % ('Image', 'Total', 'P', 'R', 'mAP'))
outputs, mAPs, mR, mP, TP, confidence, pred_class, target_class = [], [], [], [], [], [], [], [] outputs, mAPs, mR, mP, TP, confidence, pred_class, target_class = [], [], [], [], [], [], [], []
AP_accum, AP_accum_count = np.zeros(nC), np.zeros(nC) AP_accum, AP_accum_count = np.zeros(nC), np.zeros(nC)
@ -47,10 +47,10 @@ def test(
# Compute average precision for each sample # Compute average precision for each sample
for sample_i, (labels, detections) in enumerate(zip(targets, output)): for sample_i, (labels, detections) in enumerate(zip(targets, output)):
correct = [] seen += 1
if detections is None: if detections is None:
# If there are no detections but there are labels mask as zero AP # If there are labels but no detections mark as zero AP
if labels.size(0) != 0: if labels.size(0) != 0:
mAPs.append(0), mR.append(0), mP.append(0) mAPs.append(0), mR.append(0), mP.append(0)
continue continue
@ -60,6 +60,7 @@ def test(
detections = detections[np.argsort(-detections[:, 4])] detections = detections[np.argsort(-detections[:, 4])]
# If no labels add number of detections as incorrect # If no labels add number of detections as incorrect
correct = []
if labels.size(0) == 0: if labels.size(0) == 0:
# correct.extend([0 for _ in range(len(detections))]) # correct.extend([0 for _ in range(len(detections))])
mAPs.append(0), mR.append(0), mP.append(0) mAPs.append(0), mR.append(0), mP.append(0)
@ -86,7 +87,9 @@ def test(
correct.append(0) correct.append(0)
# Compute Average Precision (AP) per class # Compute Average Precision (AP) per class
AP, AP_class, R, P = ap_per_class(tp=correct, conf=detections[:, 4], pred_cls=detections[:, 6], AP, AP_class, R, P = ap_per_class(tp=correct,
conf=detections[:, 4],
pred_cls=detections[:, 6],
target_cls=target_cls) target_cls=target_cls)
# Accumulate AP per class # Accumulate AP per class
@ -104,7 +107,7 @@ def test(
mean_P = np.mean(mP) mean_P = np.mean(mP)
# Print image mAP and running mean mAP # Print image mAP and running mean mAP
print(('%11s%11s' + '%11.3g' * 3) % (len(mAPs), dataloader.nF, mean_P, mean_R, mean_mAP)) print(('%11s%11s' + '%11.3g' * 3) % (seen, dataloader.nF, mean_P, mean_R, mean_mAP))
# Print mAP per class # Print mAP per class
print('%11s' * 5 % ('Image', 'Total', 'P', 'R', 'mAP') + '\n\nmAP Per Class:') print('%11s' * 5 % ('Image', 'Total', 'P', 'R', 'mAP') + '\n\nmAP Per Class:')
@ -141,3 +144,27 @@ if __name__ == '__main__':
opt.conf_thres, opt.conf_thres,
opt.nms_thres opt.nms_thres
) )
# Image Total P R mAP # YOLOv3 320
# 32 5000 0.66 0.597 0.591
# 64 5000 0.664 0.62 0.604
# 96 5000 0.653 0.627 0.614
# 128 5000 0.639 0.623 0.607
# 160 5000 0.642 0.63 0.616
# 192 5000 0.651 0.636 0.621
# Image Total P R mAP # YOLOv3 416
# 32 5000 0.635 0.581 0.57
# 64 5000 0.63 0.591 0.578
# 96 5000 0.661 0.632 0.622
# 128 5000 0.659 0.632 0.623
# 160 5000 0.665 0.64 0.633
# 192 5000 0.66 0.637 0.63
# Image Total P R mAP # YOLOv3 608
# 32 5000 0.653 0.606 0.591
# 64 5000 0.653 0.635 0.625
# 96 5000 0.655 0.642 0.633
# 128 5000 0.667 0.651 0.642
# 160 5000 0.663 0.645 0.637
# 192 5000 0.663 0.643 0.634