garrickbrazil / m3d-rpn Goto Github PK
View Code? Open in Web Editor NEWLicense: MIT License
m3d-rpn's People
Contributors
Stargazers
Watchers
Forkers
hdjsjyl liulj13 nnu-gisa samghk qaazii spyderxu zhaott618 hyunkyulim tommyanqi fourye93 yhkim8412 big-chan jameswang007 xiaoyc003 abrahambotros marbleinc stiphyjay ansonyanxin higal tbmoon jordanyang liuguoyou jsharpclone susanmosha liufeng1990 prashantraina zechendev zhaiyuqiang aipakchoi lyx-tm collector-m riwaly oliverqian lijuanjiayou itonylee karasawatakumi motokimura xuedongji wolflion521 shipingsong pioneerge banconxuan colin131 liux9527 sykevinpeng lf2055828 senwang98 anuragsahu pamyuu fettahyildizz kim-junsung kurunie yanghh234 lzw875 g-littlesun tuskaw ted1804 elijahahianyo sunwantingsupre dashingpanda vipermdl xiazhiyi99 yinminzhang iq-scm hyewu nuaza piethonistam3d-rpn's Issues
Function missed
Hi,
It is a great work! Thank you for releasing the source code.
I find that function named projection_ray_trace and solve_transform in the file imdb_util.py are not defned. Can you fix it?
Thanks!
evlauation phase with “died with <Signals.SIGSEGV: 11>.”
After predicting labels of validation set,I use
with open(os.devnull, 'w') as devnull: out = subprocess.check_output([script, results_path.replace('/data', '')], stderr=devnull)
to load the file in path "data/kitti_split1/devkit/cpp/evaluate_object" .
However,there is an error below:
<ipython-input-5-9b215e94f907> in test_kitti_3d_back(dataset_test, test_split, rpn_conf, results_path, test_path, use_log)
----> 9 out = subprocess.check_output([script, results_path.replace('/data', '')], stderr=devnull)
10
11 for lbl in rpn_conf.lbls:
~/anaconda3/lib/python3.7/subprocess.py in check_output(timeout, *popenargs, **kwargs)
393
394 return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
--> 395 **kwargs).stdout
396
397
~/anaconda3/lib/python3.7/subprocess.py in run(input, capture_output, timeout, check, *popenargs, **kwargs)
485 if check and retcode:
486 raise CalledProcessError(retcode, process.args,
--> 487 output=stdout, stderr=stderr)
488 return CompletedProcess(process.args, retcode, stdout, stderr)
489
CalledProcessError: Command '['/home/coolguy/Project/m3d-rpn/data/kitti_split1/devkit/cpp/evaluate_object', 'output/tmp_results']' died with <Signals.SIGSEGV: 11>.
How can i fix it to get the correct evaluate results? Thanks.
error with weights after trained mode.
Hi Mr. Garrick Brazil, Thanks for your support. i trained the model. it give me test results after training finished. But when i put the same weights in the path of pretrained weights provided in this script test_rpn_3d.py. i got this error.
python3 scripts/test_rpn_3d.py
Traceback (most recent call last):
File "scripts/test_rpn_3d.py", line 49, in
load_weights(net, weights_path, remove_module=True)
File "/home/lps/M3D-RPN/lib/core.py", line 399, in load_weights
model.load_state_dict(src_weights)
File "/home/lps/.local/lib/python3.6/site-packages/torch/nn/modules/module.py", line 845, in load_state_dict
self.class.name, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for RPN:
Missing key(s) in state_dict: "cls_ble", "bbox_x_ble", "bbox_y_ble", "bbox_w_ble", "bbox_h_ble", "bbox_x3d_ble"
Can you please explain how can i test with my trained weights?
Thanks in advance.
low performance
Hi,
I am running the code to reproduce the result. I simply followed the instruction in ReadMe, everything went smoothly, but I get a low performance on both warmup and main. Here is the entire log.
Ubuntu: 16.04
Python: 3.7.7
CUDA: 10.0
Pytorch: 1.4.0
python scripts/train_rpn_3d.py --config=kitti_3d_multi_warmup
iter: 250, acc (bg: 0.97, fg: 0.34, iou: 0.60), loss (bbox_3d: 2.3948, cls: 0.7131, iou: 0.5250), misc (ry: 1.44, z: 2.85), dt: 0.51, eta: 7.1h
iter: 500, acc (bg: 0.99, fg: 0.69, iou: 0.66), loss (bbox_3d: 1.6469, cls: 0.3616, iou: 0.4295), misc (ry: 1.26, z: 2.04), dt: 0.51, eta: 7.0h
iter: 750, acc (bg: 0.99, fg: 0.79, iou: 0.70), loss (bbox_3d: 1.4113, cls: 0.2635, iou: 0.3716), misc (ry: 1.23, z: 1.84), dt: 0.51, eta: 7.0h
iter: 1000, acc (bg: 0.99, fg: 0.83, iou: 0.72), loss (bbox_3d: 1.1630, cls: 0.2295, iou: 0.3409), misc (ry: 1.05, z: 1.54), dt: 0.51, eta: 7.0h
iter: 1250, acc (bg: 0.99, fg: 0.85, iou: 0.73), loss (bbox_3d: 1.1912, cls: 0.2134, iou: 0.3225), misc (ry: 1.01, z: 1.59), dt: 0.51, eta: 6.9h
iter: 1500, acc (bg: 0.99, fg: 0.86, iou: 0.74), loss (bbox_3d: 1.0496, cls: 0.1995, iou: 0.3081), misc (ry: 0.93, z: 1.45), dt: 0.51, eta: 6.9h
iter: 1750, acc (bg: 0.99, fg: 0.85, iou: 0.75), loss (bbox_3d: 0.9926, cls: 0.2031, iou: 0.2978), misc (ry: 0.88, z: 1.41), dt: 0.51, eta: 6.9h
iter: 2000, acc (bg: 1.00, fg: 0.86, iou: 0.76), loss (bbox_3d: 0.9424, cls: 0.1804, iou: 0.2844), misc (ry: 0.90, z: 1.32), dt: 0.51, eta: 6.9h
iter: 2250, acc (bg: 0.99, fg: 0.88, iou: 0.77), loss (bbox_3d: 0.9106, cls: 0.1676, iou: 0.2699), misc (ry: 0.89, z: 1.25), dt: 0.51, eta: 6.8h
iter: 2500, acc (bg: 1.00, fg: 0.89, iou: 0.78), loss (bbox_3d: 0.7882, cls: 0.1576, iou: 0.2547), misc (ry: 0.74, z: 1.15), dt: 0.52, eta: 6.8h
iter: 2750, acc (bg: 1.00, fg: 0.90, iou: 0.79), loss (bbox_3d: 0.8344, cls: 0.1471, iou: 0.2448), misc (ry: 0.79, z: 1.28), dt: 0.52, eta: 6.8h
iter: 3000, acc (bg: 1.00, fg: 0.91, iou: 0.80), loss (bbox_3d: 0.7461, cls: 0.1343, iou: 0.2360), misc (ry: 0.76, z: 1.12), dt: 0.52, eta: 6.7h
iter: 3250, acc (bg: 1.00, fg: 0.90, iou: 0.79), loss (bbox_3d: 0.7859, cls: 0.1482, iou: 0.2374), misc (ry: 0.81, z: 1.11), dt: 0.52, eta: 6.7h
iter: 3500, acc (bg: 1.00, fg: 0.89, iou: 0.80), loss (bbox_3d: 0.7276, cls: 0.1511, iou: 0.2367), misc (ry: 0.71, z: 1.05), dt: 0.52, eta: 6.7h
iter: 3750, acc (bg: 1.00, fg: 0.89, iou: 0.79), loss (bbox_3d: 0.7858, cls: 0.1495, iou: 0.2405), misc (ry: 0.80, z: 1.11), dt: 0.52, eta: 6.6h
iter: 4000, acc (bg: 1.00, fg: 0.90, iou: 0.81), loss (bbox_3d: 0.7364, cls: 0.1357, iou: 0.2232), misc (ry: 0.68, z: 1.11), dt: 0.52, eta: 6.6h
iter: 4250, acc (bg: 1.00, fg: 0.91, iou: 0.81), loss (bbox_3d: 0.6372, cls: 0.1314, iou: 0.2174), misc (ry: 0.67, z: 0.96), dt: 0.52, eta: 6.6h
iter: 4500, acc (bg: 1.00, fg: 0.92, iou: 0.81), loss (bbox_3d: 0.6448, cls: 0.1191, iou: 0.2134), misc (ry: 0.67, z: 1.09), dt: 0.52, eta: 6.5h
iter: 4750, acc (bg: 1.00, fg: 0.91, iou: 0.82), loss (bbox_3d: 0.6515, cls: 0.1282, iou: 0.2111), misc (ry: 0.67, z: 0.96), dt: 0.52, eta: 6.5h
iter: 5000, acc (bg: 1.00, fg: 0.91, iou: 0.82), loss (bbox_3d: 0.6470, cls: 0.1284, iou: 0.2103), misc (ry: 0.70, z: 0.99), dt: 0.52, eta: 6.5h
iter: 5250, acc (bg: 1.00, fg: 0.93, iou: 0.81), loss (bbox_3d: 0.6361, cls: 0.1159, iou: 0.2146), misc (ry: 0.65, z: 1.06), dt: 0.52, eta: 6.4h
iter: 5500, acc (bg: 1.00, fg: 0.92, iou: 0.81), loss (bbox_3d: 0.6468, cls: 0.1294, iou: 0.2127), misc (ry: 0.67, z: 0.99), dt: 0.52, eta: 6.4h
iter: 5750, acc (bg: 1.00, fg: 0.91, iou: 0.82), loss (bbox_3d: 0.6072, cls: 0.1286, iou: 0.2076), misc (ry: 0.61, z: 0.95), dt: 0.52, eta: 6.3h
iter: 6000, acc (bg: 1.00, fg: 0.93, iou: 0.83), loss (bbox_3d: 0.5730, cls: 0.1137, iou: 0.1966), misc (ry: 0.62, z: 0.87), dt: 0.52, eta: 6.3h
iter: 6250, acc (bg: 1.00, fg: 0.92, iou: 0.82), loss (bbox_3d: 0.5902, cls: 0.1167, iou: 0.2004), misc (ry: 0.64, z: 0.87), dt: 0.52, eta: 6.3h
iter: 6500, acc (bg: 1.00, fg: 0.93, iou: 0.83), loss (bbox_3d: 0.5311, cls: 0.1065, iou: 0.1953), misc (ry: 0.59, z: 0.89), dt: 0.52, eta: 6.2h
iter: 6750, acc (bg: 1.00, fg: 0.92, iou: 0.83), loss (bbox_3d: 0.5250, cls: 0.1149, iou: 0.1883), misc (ry: 0.56, z: 0.88), dt: 0.52, eta: 6.2h
iter: 7000, acc (bg: 1.00, fg: 0.93, iou: 0.83), loss (bbox_3d: 0.5109, cls: 0.1076, iou: 0.1903), misc (ry: 0.51, z: 0.88), dt: 0.52, eta: 6.2h
iter: 7250, acc (bg: 1.00, fg: 0.93, iou: 0.83), loss (bbox_3d: 0.4991, cls: 0.1046, iou: 0.1887), misc (ry: 0.55, z: 0.86), dt: 0.52, eta: 6.1h
iter: 7500, acc (bg: 1.00, fg: 0.94, iou: 0.84), loss (bbox_3d: 0.4541, cls: 0.0962, iou: 0.1805), misc (ry: 0.52, z: 0.82), dt: 0.52, eta: 6.1h
iter: 7750, acc (bg: 1.00, fg: 0.93, iou: 0.84), loss (bbox_3d: 0.4874, cls: 0.1044, iou: 0.1860), misc (ry: 0.48, z: 0.90), dt: 0.52, eta: 6.1h
iter: 8000, acc (bg: 1.00, fg: 0.94, iou: 0.84), loss (bbox_3d: 0.4650, cls: 0.0982, iou: 0.1811), misc (ry: 0.51, z: 0.89), dt: 0.52, eta: 6.0h
iter: 8250, acc (bg: 1.00, fg: 0.95, iou: 0.84), loss (bbox_3d: 0.4258, cls: 0.0944, iou: 0.1768), misc (ry: 0.48, z: 0.80), dt: 0.52, eta: 6.0h
iter: 8500, acc (bg: 1.00, fg: 0.95, iou: 0.84), loss (bbox_3d: 0.4417, cls: 0.0928, iou: 0.1756), misc (ry: 0.53, z: 0.73), dt: 0.52, eta: 5.9h
iter: 8750, acc (bg: 1.00, fg: 0.96, iou: 0.85), loss (bbox_3d: 0.3894, cls: 0.0846, iou: 0.1737), misc (ry: 0.47, z: 0.78), dt: 0.52, eta: 5.9h
iter: 9000, acc (bg: 1.00, fg: 0.95, iou: 0.85), loss (bbox_3d: 0.4012, cls: 0.0891, iou: 0.1734), misc (ry: 0.46, z: 0.79), dt: 0.52, eta: 5.9h
iter: 9250, acc (bg: 1.00, fg: 0.95, iou: 0.85), loss (bbox_3d: 0.3785, cls: 0.0862, iou: 0.1723), misc (ry: 0.42, z: 0.77), dt: 0.52, eta: 5.8h
iter: 9500, acc (bg: 1.00, fg: 0.94, iou: 0.85), loss (bbox_3d: 0.4409, cls: 0.0987, iou: 0.1733), misc (ry: 0.43, z: 0.79), dt: 0.52, eta: 5.8h
iter: 9750, acc (bg: 1.00, fg: 0.95, iou: 0.84), loss (bbox_3d: 0.4245, cls: 0.0894, iou: 0.1754), misc (ry: 0.47, z: 0.78), dt: 0.52, eta: 5.8h
iter: 10000, acc (bg: 1.00, fg: 0.94, iou: 0.84), loss (bbox_3d: 0.4570, cls: 0.0964, iou: 0.1826), misc (ry: 0.48, z: 0.78), dt: 0.52, eta: 5.7h
testing 1000/3769, dt: 0.165, eta: 7.6m
testing 2000/3769, dt: 0.165, eta: 4.9m
testing 3000/3769, dt: 0.164, eta: 2.1m
test_iter 10000 2d car --> easy: 0.2364, mod: 0.2188, hard: 0.1861
test_iter 10000 gr car --> easy: 0.0274, mod: 0.0239, hard: 0.0180
test_iter 10000 3d car --> easy: 0.0187, mod: 0.0140, hard: 0.0118
test_iter 10000 2d pedestrian --> easy: 0.1260, mod: 0.1245, hard: 0.1226
test_iter 10000 gr pedestrian --> easy: 0.0116, mod: 0.0101, hard: 0.0100
test_iter 10000 3d pedestrian --> easy: 0.0065, mod: 0.0061, hard: 0.0058
test_iter 10000 2d cyclist --> easy: 0.0966, mod: 0.0951, hard: 0.0951
test_iter 10000 gr cyclist --> easy: 0.0303, mod: 0.0303, hard: 0.0303
test_iter 10000 3d cyclist --> easy: 0.0303, mod: 0.0303, hard: 0.0303
iter: 10250, acc (bg: 1.00, fg: 0.95, iou: 0.85), loss (bbox_3d: 0.3881, cls: 0.0880, iou: 0.1679), misc (ry: 0.44, z: 0.73), dt: 0.59, eta: 6.5h
iter: 10500, acc (bg: 1.00, fg: 0.95, iou: 0.85), loss (bbox_3d: 0.4284, cls: 0.0898, iou: 0.1738), misc (ry: 0.46, z: 0.83), dt: 0.58, eta: 6.4h
iter: 10750, acc (bg: 1.00, fg: 0.95, iou: 0.85), loss (bbox_3d: 0.3470, cls: 0.0857, iou: 0.1670), misc (ry: 0.40, z: 0.69), dt: 0.58, eta: 6.3h
iter: 11000, acc (bg: 1.00, fg: 0.94, iou: 0.84), loss (bbox_3d: 0.4427, cls: 0.0945, iou: 0.1759), misc (ry: 0.47, z: 0.81), dt: 0.58, eta: 6.3h
iter: 11250, acc (bg: 1.00, fg: 0.95, iou: 0.85), loss (bbox_3d: 0.4296, cls: 0.0823, iou: 0.1690), misc (ry: 0.44, z: 0.83), dt: 0.58, eta: 6.2h
iter: 11500, acc (bg: 1.00, fg: 0.96, iou: 0.85), loss (bbox_3d: 0.3720, cls: 0.0823, iou: 0.1679), misc (ry: 0.45, z: 0.73), dt: 0.58, eta: 6.2h
iter: 11750, acc (bg: 1.00, fg: 0.96, iou: 0.85), loss (bbox_3d: 0.3559, cls: 0.0792, iou: 0.1614), misc (ry: 0.48, z: 0.70), dt: 0.58, eta: 6.1h
iter: 12000, acc (bg: 1.00, fg: 0.95, iou: 0.85), loss (bbox_3d: 0.3837, cls: 0.0883, iou: 0.1679), misc (ry: 0.42, z: 0.74), dt: 0.57, eta: 6.1h
iter: 12250, acc (bg: 1.00, fg: 0.95, iou: 0.86), loss (bbox_3d: 0.3504, cls: 0.0866, iou: 0.1611), misc (ry: 0.42, z: 0.73), dt: 0.57, eta: 6.0h
iter: 12500, acc (bg: 1.00, fg: 0.95, iou: 0.84), loss (bbox_3d: 0.4260, cls: 0.0917, iou: 0.1770), misc (ry: 0.52, z: 0.74), dt: 0.57, eta: 6.0h
iter: 12750, acc (bg: 1.00, fg: 0.95, iou: 0.86), loss (bbox_3d: 0.3284, cls: 0.0831, iou: 0.1610), misc (ry: 0.40, z: 0.71), dt: 0.57, eta: 5.9h
iter: 13000, acc (bg: 1.00, fg: 0.95, iou: 0.86), loss (bbox_3d: 0.3203, cls: 0.0822, iou: 0.1594), misc (ry: 0.36, z: 0.66), dt: 0.57, eta: 5.8h
iter: 13250, acc (bg: 1.00, fg: 0.96, iou: 0.86), loss (bbox_3d: 0.3265, cls: 0.0772, iou: 0.1567), misc (ry: 0.36, z: 0.70), dt: 0.57, eta: 5.8h
iter: 13500, acc (bg: 1.00, fg: 0.95, iou: 0.86), loss (bbox_3d: 0.3445, cls: 0.0804, iou: 0.1610), misc (ry: 0.44, z: 0.68), dt: 0.57, eta: 5.7h
iter: 13750, acc (bg: 1.00, fg: 0.96, iou: 0.86), loss (bbox_3d: 0.2789, cls: 0.0723, iou: 0.1504), misc (ry: 0.33, z: 0.62), dt: 0.57, eta: 5.7h
iter: 14000, acc (bg: 1.00, fg: 0.96, iou: 0.87), loss (bbox_3d: 0.3030, cls: 0.0721, iou: 0.1481), misc (ry: 0.40, z: 0.63), dt: 0.56, eta: 5.6h
iter: 14250, acc (bg: 1.00, fg: 0.96, iou: 0.85), loss (bbox_3d: 0.3612, cls: 0.0782, iou: 0.1656), misc (ry: 0.46, z: 0.73), dt: 0.56, eta: 5.6h
iter: 14500, acc (bg: 1.00, fg: 0.96, iou: 0.86), loss (bbox_3d: 0.3331, cls: 0.0761, iou: 0.1549), misc (ry: 0.41, z: 0.61), dt: 0.56, eta: 5.6h
iter: 14750, acc (bg: 1.00, fg: 0.96, iou: 0.86), loss (bbox_3d: 0.3069, cls: 0.0713, iou: 0.1565), misc (ry: 0.39, z: 0.67), dt: 0.56, eta: 5.5h
iter: 15000, acc (bg: 1.00, fg: 0.96, iou: 0.86), loss (bbox_3d: 0.2793, cls: 0.0739, iou: 0.1502), misc (ry: 0.35, z: 0.61), dt: 0.56, eta: 5.5h
iter: 15250, acc (bg: 1.00, fg: 0.96, iou: 0.87), loss (bbox_3d: 0.3163, cls: 0.0744, iou: 0.1487), misc (ry: 0.40, z: 0.63), dt: 0.56, eta: 5.4h
iter: 15500, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2504, cls: 0.0663, iou: 0.1397), misc (ry: 0.34, z: 0.62), dt: 0.56, eta: 5.4h
iter: 15750, acc (bg: 1.00, fg: 0.96, iou: 0.87), loss (bbox_3d: 0.2597, cls: 0.0712, iou: 0.1474), misc (ry: 0.31, z: 0.60), dt: 0.56, eta: 5.3h
iter: 16000, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2488, cls: 0.0657, iou: 0.1423), misc (ry: 0.30, z: 0.59), dt: 0.56, eta: 5.3h
iter: 16250, acc (bg: 1.00, fg: 0.96, iou: 0.87), loss (bbox_3d: 0.2739, cls: 0.0690, iou: 0.1447), misc (ry: 0.34, z: 0.66), dt: 0.56, eta: 5.2h
iter: 16500, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2521, cls: 0.0634, iou: 0.1378), misc (ry: 0.36, z: 0.60), dt: 0.56, eta: 5.2h
iter: 16750, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2214, cls: 0.0629, iou: 0.1373), misc (ry: 0.31, z: 0.59), dt: 0.56, eta: 5.1h
iter: 17000, acc (bg: 1.00, fg: 0.96, iou: 0.87), loss (bbox_3d: 0.2629, cls: 0.0688, iou: 0.1431), misc (ry: 0.36, z: 0.59), dt: 0.55, eta: 5.1h
iter: 17250, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2066, cls: 0.0632, iou: 0.1333), misc (ry: 0.28, z: 0.52), dt: 0.55, eta: 5.0h
iter: 17500, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2392, cls: 0.0624, iou: 0.1389), misc (ry: 0.32, z: 0.57), dt: 0.55, eta: 5.0h
iter: 17750, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2565, cls: 0.0614, iou: 0.1365), misc (ry: 0.33, z: 0.66), dt: 0.55, eta: 5.0h
iter: 18000, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2476, cls: 0.0655, iou: 0.1394), misc (ry: 0.32, z: 0.64), dt: 0.55, eta: 4.9h
iter: 18250, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2328, cls: 0.0652, iou: 0.1374), misc (ry: 0.32, z: 0.60), dt: 0.55, eta: 4.9h
iter: 18500, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2271, cls: 0.0622, iou: 0.1369), misc (ry: 0.29, z: 0.61), dt: 0.55, eta: 4.8h
iter: 18750, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2322, cls: 0.0606, iou: 0.1336), misc (ry: 0.29, z: 0.59), dt: 0.55, eta: 4.8h
iter: 19000, acc (bg: 1.00, fg: 0.98, iou: 0.88), loss (bbox_3d: 0.2134, cls: 0.0590, iou: 0.1328), misc (ry: 0.30, z: 0.56), dt: 0.55, eta: 4.7h
iter: 19250, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2441, cls: 0.0630, iou: 0.1396), misc (ry: 0.34, z: 0.62), dt: 0.55, eta: 4.7h
iter: 19500, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2222, cls: 0.0639, iou: 0.1349), misc (ry: 0.30, z: 0.59), dt: 0.55, eta: 4.7h
iter: 19750, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2459, cls: 0.0585, iou: 0.1349), misc (ry: 0.36, z: 0.57), dt: 0.55, eta: 4.6h
iter: 20000, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2246, cls: 0.0630, iou: 0.1360), misc (ry: 0.29, z: 0.59), dt: 0.55, eta: 4.6h
testing 1000/3769, dt: 0.164, eta: 7.6m
testing 2000/3769, dt: 0.163, eta: 4.8m
testing 3000/3769, dt: 0.163, eta: 2.1m
test_iter 20000 2d car --> easy: 0.1632, mod: 0.2028, hard: 0.1749
test_iter 20000 gr car --> easy: 0.0344, mod: 0.0341, hard: 0.0275
test_iter 20000 3d car --> easy: 0.0250, mod: 0.0239, hard: 0.0199
test_iter 20000 2d pedestrian --> easy: 0.1343, mod: 0.1255, hard: 0.1186
test_iter 20000 gr pedestrian --> easy: 0.0179, mod: 0.0178, hard: 0.0156
test_iter 20000 3d pedestrian --> easy: 0.0168, mod: 0.0151, hard: 0.0151
test_iter 20000 2d cyclist --> easy: 0.0215, mod: 0.0208, hard: 0.0209
test_iter 20000 gr cyclist --> easy: 0.0010, mod: 0.0011, hard: 0.0014
test_iter 20000 3d cyclist --> easy: 0.0010, mod: 0.0010, hard: 0.0010
iter: 20250, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.1996, cls: 0.0589, iou: 0.1289), misc (ry: 0.28, z: 0.54), dt: 0.58, eta: 4.8h
iter: 20500, acc (bg: 1.00, fg: 0.98, iou: 0.88), loss (bbox_3d: 0.2049, cls: 0.0575, iou: 0.1307), misc (ry: 0.32, z: 0.54), dt: 0.58, eta: 4.8h
iter: 20750, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2004, cls: 0.0588, iou: 0.1302), misc (ry: 0.28, z: 0.52), dt: 0.58, eta: 4.7h
iter: 21000, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2035, cls: 0.0588, iou: 0.1271), misc (ry: 0.31, z: 0.57), dt: 0.58, eta: 4.7h
iter: 21250, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.1917, cls: 0.0565, iou: 0.1278), misc (ry: 0.31, z: 0.49), dt: 0.58, eta: 4.6h
iter: 21500, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1836, cls: 0.0532, iou: 0.1223), misc (ry: 0.29, z: 0.49), dt: 0.58, eta: 4.6h
iter: 21750, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2027, cls: 0.0574, iou: 0.1297), misc (ry: 0.29, z: 0.55), dt: 0.58, eta: 4.5h
iter: 22000, acc (bg: 1.00, fg: 0.98, iou: 0.88), loss (bbox_3d: 0.1927, cls: 0.0561, iou: 0.1270), misc (ry: 0.26, z: 0.54), dt: 0.58, eta: 4.5h
iter: 22250, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2391, cls: 0.0600, iou: 0.1324), misc (ry: 0.37, z: 0.54), dt: 0.58, eta: 4.4h
iter: 22500, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1988, cls: 0.0547, iou: 0.1248), misc (ry: 0.33, z: 0.51), dt: 0.58, eta: 4.4h
iter: 22750, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.2084, cls: 0.0567, iou: 0.1238), misc (ry: 0.32, z: 0.52), dt: 0.57, eta: 4.4h
iter: 23000, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.1843, cls: 0.0591, iou: 0.1268), misc (ry: 0.28, z: 0.49), dt: 0.57, eta: 4.3h
iter: 23250, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1611, cls: 0.0513, iou: 0.1201), misc (ry: 0.27, z: 0.47), dt: 0.57, eta: 4.3h
iter: 23500, acc (bg: 1.00, fg: 0.98, iou: 0.88), loss (bbox_3d: 0.1670, cls: 0.0518, iou: 0.1262), misc (ry: 0.25, z: 0.47), dt: 0.57, eta: 4.2h
iter: 23750, acc (bg: 1.00, fg: 0.97, iou: 0.89), loss (bbox_3d: 0.1711, cls: 0.0563, iou: 0.1239), misc (ry: 0.27, z: 0.50), dt: 0.57, eta: 4.2h
iter: 24000, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1863, cls: 0.0543, iou: 0.1236), misc (ry: 0.30, z: 0.50), dt: 0.57, eta: 4.1h
iter: 24250, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1818, cls: 0.0538, iou: 0.1226), misc (ry: 0.31, z: 0.49), dt: 0.57, eta: 4.1h
iter: 24500, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1793, cls: 0.0521, iou: 0.1170), misc (ry: 0.32, z: 0.48), dt: 0.57, eta: 4.0h
iter: 24750, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1798, cls: 0.0532, iou: 0.1199), misc (ry: 0.28, z: 0.50), dt: 0.57, eta: 4.0h
iter: 25000, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1587, cls: 0.0511, iou: 0.1172), misc (ry: 0.27, z: 0.48), dt: 0.57, eta: 3.9h
iter: 25250, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1544, cls: 0.0536, iou: 0.1172), misc (ry: 0.22, z: 0.47), dt: 0.57, eta: 3.9h
iter: 25500, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1465, cls: 0.0502, iou: 0.1146), misc (ry: 0.24, z: 0.46), dt: 0.57, eta: 3.9h
iter: 25750, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1753, cls: 0.0520, iou: 0.1188), misc (ry: 0.28, z: 0.49), dt: 0.57, eta: 3.8h
iter: 26000, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1482, cls: 0.0491, iou: 0.1159), misc (ry: 0.22, z: 0.48), dt: 0.57, eta: 3.8h
iter: 26250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1305, cls: 0.0463, iou: 0.1089), misc (ry: 0.22, z: 0.43), dt: 0.57, eta: 3.7h
iter: 26500, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1473, cls: 0.0516, iou: 0.1171), misc (ry: 0.24, z: 0.47), dt: 0.57, eta: 3.7h
iter: 26750, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1395, cls: 0.0501, iou: 0.1086), misc (ry: 0.22, z: 0.44), dt: 0.56, eta: 3.6h
iter: 27000, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1374, cls: 0.0481, iou: 0.1098), misc (ry: 0.22, z: 0.47), dt: 0.56, eta: 3.6h
iter: 27250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1309, cls: 0.0502, iou: 0.1087), misc (ry: 0.20, z: 0.43), dt: 0.56, eta: 3.6h
iter: 27500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1357, cls: 0.0477, iou: 0.1123), misc (ry: 0.23, z: 0.44), dt: 0.56, eta: 3.5h
iter: 27750, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1244, cls: 0.0479, iou: 0.1064), misc (ry: 0.21, z: 0.43), dt: 0.56, eta: 3.5h
iter: 28000, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1184, cls: 0.0453, iou: 0.1046), misc (ry: 0.21, z: 0.41), dt: 0.56, eta: 3.4h
iter: 28250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1332, cls: 0.0475, iou: 0.1099), misc (ry: 0.23, z: 0.41), dt: 0.56, eta: 3.4h
iter: 28500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1290, cls: 0.0486, iou: 0.1112), misc (ry: 0.21, z: 0.46), dt: 0.56, eta: 3.4h
iter: 28750, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1212, cls: 0.0463, iou: 0.1051), misc (ry: 0.21, z: 0.44), dt: 0.56, eta: 3.3h
iter: 29000, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1441, cls: 0.0501, iou: 0.1111), misc (ry: 0.21, z: 0.49), dt: 0.56, eta: 3.3h
iter: 29250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1254, cls: 0.0449, iou: 0.1062), misc (ry: 0.21, z: 0.44), dt: 0.56, eta: 3.2h
iter: 29500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1139, cls: 0.0458, iou: 0.1029), misc (ry: 0.20, z: 0.43), dt: 0.56, eta: 3.2h
iter: 29750, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1152, cls: 0.0423, iou: 0.1007), misc (ry: 0.19, z: 0.40), dt: 0.56, eta: 3.1h
iter: 30000, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1262, cls: 0.0486, iou: 0.1068), misc (ry: 0.21, z: 0.41), dt: 0.56, eta: 3.1h
testing 1000/3769, dt: 0.162, eta: 7.5m
testing 2000/3769, dt: 0.162, eta: 4.8m
testing 3000/3769, dt: 0.162, eta: 2.1m
test_iter 30000 2d car --> easy: 0.1578, mod: 0.1938, hard: 0.1675
test_iter 30000 gr car --> easy: 0.0845, mod: 0.1187, hard: 0.1115
test_iter 30000 3d car --> easy: 0.0752, mod: 0.1110, hard: 0.1049
test_iter 30000 2d pedestrian --> easy: 0.1269, mod: 0.1216, hard: 0.1151
test_iter 30000 gr pedestrian --> easy: 0.0164, mod: 0.0290, hard: 0.0284
test_iter 30000 3d pedestrian --> easy: 0.0121, mod: 0.0281, hard: 0.0256
test_iter 30000 2d cyclist --> easy: 0.0986, mod: 0.0948, hard: 0.0949
test_iter 30000 gr cyclist --> easy: 0.0124, mod: 0.0101, hard: 0.0101
test_iter 30000 3d cyclist --> easy: 0.0118, mod: 0.0101, hard: 0.0101
iter: 30250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1275, cls: 0.0467, iou: 0.1045), misc (ry: 0.22, z: 0.44), dt: 0.58, eta: 3.2h
iter: 30500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1241, cls: 0.0453, iou: 0.1050), misc (ry: 0.19, z: 0.48), dt: 0.58, eta: 3.1h
iter: 30750, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1198, cls: 0.0434, iou: 0.1016), misc (ry: 0.19, z: 0.42), dt: 0.58, eta: 3.1h
iter: 31000, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.0979, cls: 0.0421, iou: 0.0954), misc (ry: 0.18, z: 0.40), dt: 0.58, eta: 3.1h
iter: 31250, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1063, cls: 0.0437, iou: 0.1012), misc (ry: 0.20, z: 0.39), dt: 0.58, eta: 3.0h
iter: 31500, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1053, cls: 0.0446, iou: 0.0999), misc (ry: 0.19, z: 0.39), dt: 0.58, eta: 3.0h
iter: 31750, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1068, cls: 0.0419, iou: 0.0981), misc (ry: 0.19, z: 0.40), dt: 0.58, eta: 2.9h
iter: 32000, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.1066, cls: 0.0417, iou: 0.0968), misc (ry: 0.19, z: 0.41), dt: 0.58, eta: 2.9h
iter: 32250, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.1059, cls: 0.0418, iou: 0.0984), misc (ry: 0.19, z: 0.39), dt: 0.58, eta: 2.8h
iter: 32500, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.0946, cls: 0.0436, iou: 0.0941), misc (ry: 0.18, z: 0.36), dt: 0.58, eta: 2.8h
iter: 32750, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.0929, cls: 0.0447, iou: 0.0959), misc (ry: 0.18, z: 0.36), dt: 0.58, eta: 2.8h
iter: 33000, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.0978, cls: 0.0404, iou: 0.0970), misc (ry: 0.18, z: 0.40), dt: 0.58, eta: 2.7h
iter: 33250, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.1015, cls: 0.0402, iou: 0.0962), misc (ry: 0.18, z: 0.37), dt: 0.57, eta: 2.7h
iter: 33500, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1100, cls: 0.0420, iou: 0.0932), misc (ry: 0.22, z: 0.35), dt: 0.57, eta: 2.6h
iter: 33750, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1134, cls: 0.0435, iou: 0.0979), misc (ry: 0.22, z: 0.38), dt: 0.57, eta: 2.6h
iter: 34000, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1128, cls: 0.0422, iou: 0.0965), misc (ry: 0.22, z: 0.40), dt: 0.57, eta: 2.5h
iter: 34250, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.0975, cls: 0.0398, iou: 0.0930), misc (ry: 0.18, z: 0.38), dt: 0.57, eta: 2.5h
iter: 34500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0867, cls: 0.0398, iou: 0.0887), misc (ry: 0.17, z: 0.36), dt: 0.57, eta: 2.5h
iter: 34750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0828, cls: 0.0388, iou: 0.0881), misc (ry: 0.17, z: 0.34), dt: 0.57, eta: 2.4h
iter: 35000, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.0859, cls: 0.0400, iou: 0.0915), misc (ry: 0.17, z: 0.35), dt: 0.57, eta: 2.4h
iter: 35250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0870, cls: 0.0404, iou: 0.0864), misc (ry: 0.18, z: 0.33), dt: 0.57, eta: 2.3h
iter: 35500, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.0906, cls: 0.0399, iou: 0.0936), misc (ry: 0.18, z: 0.36), dt: 0.57, eta: 2.3h
iter: 35750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0821, cls: 0.0400, iou: 0.0867), misc (ry: 0.17, z: 0.35), dt: 0.57, eta: 2.3h
iter: 36000, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.0921, cls: 0.0428, iou: 0.0927), misc (ry: 0.18, z: 0.37), dt: 0.57, eta: 2.2h
iter: 36250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0798, cls: 0.0389, iou: 0.0880), misc (ry: 0.16, z: 0.33), dt: 0.57, eta: 2.2h
iter: 36500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0744, cls: 0.0352, iou: 0.0833), misc (ry: 0.15, z: 0.35), dt: 0.57, eta: 2.1h
iter: 36750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0776, cls: 0.0390, iou: 0.0870), misc (ry: 0.15, z: 0.33), dt: 0.57, eta: 2.1h
iter: 37000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0705, cls: 0.0366, iou: 0.0816), misc (ry: 0.15, z: 0.33), dt: 0.57, eta: 2.1h
iter: 37250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0877, cls: 0.0405, iou: 0.0888), misc (ry: 0.17, z: 0.33), dt: 0.57, eta: 2.0h
iter: 37500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0660, cls: 0.0374, iou: 0.0819), misc (ry: 0.15, z: 0.30), dt: 0.57, eta: 2.0h
iter: 37750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0750, cls: 0.0364, iou: 0.0819), misc (ry: 0.16, z: 0.33), dt: 0.57, eta: 1.9h
iter: 38000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0724, cls: 0.0342, iou: 0.0794), misc (ry: 0.15, z: 0.34), dt: 0.57, eta: 1.9h
iter: 38250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0738, cls: 0.0371, iou: 0.0841), misc (ry: 0.16, z: 0.32), dt: 0.57, eta: 1.8h
iter: 38500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0713, cls: 0.0370, iou: 0.0808), misc (ry: 0.15, z: 0.32), dt: 0.57, eta: 1.8h
iter: 38750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0706, cls: 0.0353, iou: 0.0806), misc (ry: 0.15, z: 0.31), dt: 0.57, eta: 1.8h
iter: 39000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0742, cls: 0.0390, iou: 0.0842), misc (ry: 0.15, z: 0.31), dt: 0.57, eta: 1.7h
iter: 39250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0693, cls: 0.0380, iou: 0.0821), misc (ry: 0.15, z: 0.30), dt: 0.56, eta: 1.7h
iter: 39500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0692, cls: 0.0362, iou: 0.0800), misc (ry: 0.15, z: 0.31), dt: 0.56, eta: 1.6h
iter: 39750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0731, cls: 0.0360, iou: 0.0817), misc (ry: 0.15, z: 0.32), dt: 0.56, eta: 1.6h
iter: 40000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0675, cls: 0.0344, iou: 0.0806), misc (ry: 0.14, z: 0.33), dt: 0.56, eta: 1.6h
testing 1000/3769, dt: 0.160, eta: 7.4m
testing 2000/3769, dt: 0.160, eta: 4.7m
testing 3000/3769, dt: 0.160, eta: 2.1m
test_iter 40000 2d car --> easy: 0.1608, mod: 0.1554, hard: 0.1375
test_iter 40000 gr car --> easy: 0.1229, mod: 0.1147, hard: 0.1081
test_iter 40000 3d car --> easy: 0.1159, mod: 0.1080, hard: 0.1027
test_iter 40000 2d pedestrian --> easy: 0.1353, mod: 0.1287, hard: 0.1214
test_iter 40000 gr pedestrian --> easy: 0.0236, mod: 0.0990, hard: 0.0975
test_iter 40000 3d pedestrian --> easy: 0.0217, mod: 0.0980, hard: 0.0965
test_iter 40000 2d cyclist --> easy: 0.0996, mod: 0.0945, hard: 0.0944
test_iter 40000 gr cyclist --> easy: 0.0049, mod: 0.0035, hard: 0.0035
test_iter 40000 3d cyclist --> easy: 0.0049, mod: 0.0035, hard: 0.0035
iter: 40250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0624, cls: 0.0330, iou: 0.0775), misc (ry: 0.14, z: 0.32), dt: 0.58, eta: 1.6h
iter: 40500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0639, cls: 0.0357, iou: 0.0786), misc (ry: 0.14, z: 0.31), dt: 0.58, eta: 1.5h
iter: 40750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0609, cls: 0.0347, iou: 0.0757), misc (ry: 0.14, z: 0.29), dt: 0.58, eta: 1.5h
iter: 41000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0595, cls: 0.0359, iou: 0.0762), misc (ry: 0.14, z: 0.29), dt: 0.58, eta: 1.4h
iter: 41250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0615, cls: 0.0353, iou: 0.0751), misc (ry: 0.15, z: 0.29), dt: 0.58, eta: 1.4h
iter: 41500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0648, cls: 0.0358, iou: 0.0795), misc (ry: 0.14, z: 0.31), dt: 0.58, eta: 1.4h
iter: 41750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0592, cls: 0.0351, iou: 0.0739), misc (ry: 0.14, z: 0.29), dt: 0.58, eta: 1.3h
iter: 42000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0560, cls: 0.0341, iou: 0.0732), misc (ry: 0.13, z: 0.28), dt: 0.58, eta: 1.3h
iter: 42250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0591, cls: 0.0346, iou: 0.0745), misc (ry: 0.14, z: 0.29), dt: 0.58, eta: 1.2h
iter: 42500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0573, cls: 0.0360, iou: 0.0733), misc (ry: 0.13, z: 0.29), dt: 0.58, eta: 1.2h
iter: 42750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0551, cls: 0.0331, iou: 0.0711), misc (ry: 0.14, z: 0.28), dt: 0.58, eta: 1.2h
iter: 43000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0607, cls: 0.0350, iou: 0.0719), misc (ry: 0.14, z: 0.28), dt: 0.58, eta: 1.1h
iter: 43250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0531, cls: 0.0332, iou: 0.0703), misc (ry: 0.13, z: 0.29), dt: 0.58, eta: 1.1h
iter: 43500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0517, cls: 0.0324, iou: 0.0717), misc (ry: 0.13, z: 0.28), dt: 0.58, eta: 1.0h
iter: 43750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0552, cls: 0.0337, iou: 0.0693), misc (ry: 0.13, z: 0.28), dt: 0.57, eta: 59.9m
iter: 44000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0614, cls: 0.0354, iou: 0.0757), misc (ry: 0.14, z: 0.30), dt: 0.57, eta: 57.4m
iter: 44250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0472, cls: 0.0325, iou: 0.0669), misc (ry: 0.12, z: 0.26), dt: 0.57, eta: 55.0m
iter: 44500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0540, cls: 0.0336, iou: 0.0691), misc (ry: 0.13, z: 0.28), dt: 0.57, eta: 52.6m
iter: 44750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0533, cls: 0.0328, iou: 0.0675), misc (ry: 0.13, z: 0.27), dt: 0.57, eta: 50.2m
iter: 45000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0498, cls: 0.0326, iou: 0.0677), misc (ry: 0.13, z: 0.27), dt: 0.57, eta: 47.7m
iter: 45250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0517, cls: 0.0338, iou: 0.0683), misc (ry: 0.13, z: 0.27), dt: 0.57, eta: 45.3m
iter: 45500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0470, cls: 0.0305, iou: 0.0658), misc (ry: 0.12, z: 0.27), dt: 0.57, eta: 42.9m
iter: 45750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0478, cls: 0.0313, iou: 0.0651), misc (ry: 0.12, z: 0.26), dt: 0.57, eta: 40.5m
iter: 46000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0505, cls: 0.0316, iou: 0.0659), misc (ry: 0.12, z: 0.27), dt: 0.57, eta: 38.1m
iter: 46250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0450, cls: 0.0305, iou: 0.0618), misc (ry: 0.12, z: 0.26), dt: 0.57, eta: 35.7m
iter: 46500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0524, cls: 0.0340, iou: 0.0665), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 33.3m
iter: 46750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0477, cls: 0.0333, iou: 0.0644), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 30.9m
iter: 47000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0531, cls: 0.0342, iou: 0.0680), misc (ry: 0.13, z: 0.27), dt: 0.57, eta: 28.5m
iter: 47250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0454, cls: 0.0303, iou: 0.0630), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 26.1m
iter: 47500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0464, cls: 0.0324, iou: 0.0623), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 23.7m
iter: 47750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0476, cls: 0.0328, iou: 0.0652), misc (ry: 0.12, z: 0.26), dt: 0.57, eta: 21.3m
iter: 48000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0536, cls: 0.0349, iou: 0.0666), misc (ry: 0.13, z: 0.25), dt: 0.57, eta: 19.0m
iter: 48250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0477, cls: 0.0323, iou: 0.0640), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 16.6m
iter: 48500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0469, cls: 0.0306, iou: 0.0627), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 14.2m
iter: 48750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0448, cls: 0.0320, iou: 0.0623), misc (ry: 0.12, z: 0.24), dt: 0.57, eta: 11.8m
iter: 49000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0466, cls: 0.0307, iou: 0.0614), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 9.5m
iter: 49250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0466, cls: 0.0325, iou: 0.0620), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 7.1m
iter: 49500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0483, cls: 0.0332, iou: 0.0628), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 4.7m
iter: 49750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0437, cls: 0.0319, iou: 0.0605), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 2.4m
iter: 50000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0484, cls: 0.0338, iou: 0.0627), misc (ry: 0.12, z: 0.25), dt: 0.57, eta: 0.6s
testing 1000/3769, dt: 0.159, eta: 7.3m
testing 2000/3769, dt: 0.159, eta: 4.7m
testing 3000/3769, dt: 0.159, eta: 2.0m
test_iter 50000 2d car --> easy: 0.1439, mod: 0.1443, hard: 0.1296
test_iter 50000 gr car --> easy: 0.1152, mod: 0.1092, hard: 0.1063
test_iter 50000 3d car --> easy: 0.1109, mod: 0.1061, hard: 0.1017
test_iter 50000 2d pedestrian --> easy: 0.1349, mod: 0.1286, hard: 0.1215
test_iter 50000 gr pedestrian --> easy: 0.0412, mod: 0.0986, hard: 0.0975
test_iter 50000 3d pedestrian --> easy: 0.0402, mod: 0.0976, hard: 0.0970
test_iter 50000 2d cyclist --> easy: 0.1028, mod: 0.0968, hard: 0.0967
test_iter 50000 gr cyclist --> easy: 0.0484, mod: 0.0477, hard: 0.0455
test_iter 50000 3d cyclist --> easy: 0.0480, mod: 0.0455, hard: 0.0455
Segmentation fault (core dumped)
python scripts/train_rpn_3d.py --config=kitti_3d_multi_main
Computing bbox regression mean..
Computing bbox regression stds..
used 814041 boxes with avg std 0.6826
conf: {
model: densenet121_3d_dilate_depth_aware
solver_type: sgd
lr: 0.004
momentum: 0.9
weight_decay: 0.0005
max_iter: 50000
snapshot_iter: 10000
display: 250
do_test: True
lr_policy: poly
lr_steps: None
lr_target: 4e-08
rng_seed: 2
cuda_seed: 2
image_means: [0.485, 0.456, 0.406]
image_stds: [0.229, 0.224, 0.225]
feat_stride: 16
has_3d: True
test_scale: 512
crop_size: [512, 1760]
mirror_prob: 0.5
distort_prob: -1
dataset_test: kitti_split1
datasets_train: [{'anno_fmt': 'kitti_det',
'im_ext': '.png',
'name': 'kitti_split1',
'scale': 1}]
use_3d_for_2d: True
percent_anc_h: [0.0625, 0.75]
min_gt_h: 32.0
max_gt_h: 384.0
min_gt_vis: 0.65
ilbls: ['Van', 'ignore']
lbls: ['Car', 'Pedestrian', 'Cyclist']
batch_size: 2
fg_image_ratio: 1.0
box_samples: 0.2
fg_fraction: 0.2
bg_thresh_lo: 0
bg_thresh_hi: 0.5
fg_thresh: 0.5
ign_thresh: 0.5
best_thresh: 0.35
nms_topN_pre: 3000
nms_topN_post: 40
nms_thres: 0.4
clip_boxes: False
test_protocol: kitti
test_db: kitti
test_min_h: 0
min_det_scales: [0, 0]
cluster_anchors: 0
even_anchors: 0
expand_anchors: 0
anchors: [[-0.5, -8.5, 15.5, 23.5, 51.969, 0.531,
1.713, 1.025, -0.799],
[-8.5, -8.5, 23.5, 23.5, 52.176, 1.618,
1.6, 3.811, -0.453],
[-16.5, -8.5, 31.5, 23.5, 48.334,
1.644, 1.529, 3.966, 0.673],
[-2.528, -12.555, 17.528, 27.555,
44.781, 0.534, 1.771, 0.971, 0.093],
[-12.555, -12.555, 27.555, 27.555,
44.704, 1.599, 1.569, 3.814, -0.187],
[-22.583, -12.555, 37.583, 27.555,
43.492, 1.621, 1.536, 3.91, 0.719],
[-5.069, -17.638, 20.069, 32.638,
34.666, 0.561, 1.752, 0.967, -0.384],
[-17.638, -17.638, 32.638, 32.638,
35.35, 1.567, 1.591, 3.81, -0.511],
[-30.207, -17.638, 45.207, 32.638,
37.128, 1.602, 1.529, 3.904, 0.452],
[-8.255, -24.01, 23.255, 39.01, 28.771,
0.613, 1.76, 0.98, 0.067],
[-24.01, -24.01, 39.01, 39.01, 28.331,
1.543, 1.592, 3.66, -0.811],
[-39.764, -24.01, 54.764, 39.01,
30.541, 1.626, 1.524, 3.908, 0.312],
[-12.248, -31.996, 27.248, 46.996,
23.011, 0.606, 1.758, 0.996, 0.208],
[-31.996, -31.996, 46.996, 46.996,
22.948, 1.51, 1.599, 3.419, -1.076],
[-51.744, -31.996, 66.744, 46.996,
25.0, 1.628, 1.527, 3.917, 0.334],
[-17.253, -42.006, 32.253, 57.006,
18.479, 0.601, 1.747, 1.007, 0.347],
[-42.006, -42.006, 57.006, 57.006,
18.815, 1.487, 1.599, 3.337, -0.862],
[-66.759, -42.006, 81.759, 57.006,
20.576, 1.623, 1.532, 3.942, 0.323],
[-23.527, -54.553, 38.527, 69.553,
15.035, 0.625, 1.744, 0.917, 0.41],
[-54.553, -54.553, 69.553, 69.553,
15.346, 1.29, 1.659, 3.083, -0.275],
[-85.58, -54.553, 100.58, 69.553,
16.326, 1.613, 1.527, 3.934, 0.268],
[-31.39, -70.281, 46.39, 85.281,
12.265, 0.631, 1.747, 0.954, 0.317],
[-70.281, -70.281, 85.281, 85.281,
11.878, 1.044, 1.67, 2.415, -0.211],
[-109.171, -70.281, 124.171, 85.281,
13.58, 1.621, 1.539, 3.961, 0.189],
[-41.247, -89.994, 56.247, 104.994,
9.932, 0.61, 1.771, 0.934, 0.486],
[-89.994, -89.994, 104.994, 104.994,
8.949, 0.811, 1.766, 1.662, 0.08],
[-138.741, -89.994, 153.741, 104.994,
11.043, 1.61, 1.533, 3.899, 0.04],
[-53.602, -114.704, 68.602, 129.704,
8.389, 0.604, 1.793, 0.95, 0.806],
[-114.704, -114.704, 129.704, 129.704,
8.071, 1.01, 1.751, 2.19, -0.076],
[-175.806, -114.704, 190.806, 129.704,
9.184, 1.606, 1.526, 3.869, -0.066],
[-69.089, -145.677, 84.089, 160.677,
6.923, 0.627, 1.791, 0.96, 0.784],
[-145.677, -145.677, 160.677, 160.677,
6.784, 1.384, 1.615, 2.862, -1.035],
[-222.266, -145.677, 237.266, 160.677,
7.863, 1.617, 1.55, 3.948, -0.071],
[-88.5, -184.5, 103.5, 199.5, 5.189,
0.66, 1.755, 0.841, 0.173],
[-184.5, -184.5, 199.5, 199.5, 4.388,
0.743, 1.728, 1.381, 0.642],
[-280.5, -184.5, 295.5, 199.5, 5.583,
1.583, 1.547, 3.862, -0.072]]
bbox_means: [[-0.0, 0.002, 0.064, -0.093, 0.011,
-0.067, 0.192, 0.059, -0.021, 0.069,
-0.004]]
bbox_stds: [[0.14, 0.126, 0.247, 0.239, 0.163,
0.132, 3.621, 0.382, 0.102, 0.503,
1.855]]
anchor_scales: [32.0, 40.11, 50.276, 63.019, 78.991,
99.012, 124.106, 155.561, 194.989,
244.409, 306.354, 384.0]
anchor_ratios: [0.5, 1.0, 1.5]
hard_negatives: True
focal_loss: 0
cls_2d_lambda: 1
iou_2d_lambda: 1
bbox_2d_lambda: 0
bbox_3d_lambda: 1
bbox_3d_proj_lambda: 0.0
hill_climbing: True
bins: 32
visdom_port: 8100
pretrained: 'output/kitti_3d_multi_warmup/weights/model_50000_pkl'
}
/home/robot1/anaconda3/envs/M3DRPN/lib/python3.6/site-packages/torchvision/models/densenet.py:212: UserWarning: nn.init.kaiming_normal is now deprecated in favor of nn.init.kaiming_normal_.
nn.init.kaiming_normal(m.weight.data)
iter: 250, acc (bg: 1.00, fg: 0.97, iou: 0.90), loss (bbox_3d: 0.1411, cls: 0.0445, iou: 0.1096), misc (ry: 0.21, z: 0.49), dt: 0.60, eta: 8.3h
iter: 500, acc (bg: 1.00, fg: 0.94, iou: 0.85), loss (bbox_3d: 0.4028, cls: 0.0901, iou: 0.1682), misc (ry: 0.44, z: 0.79), dt: 0.60, eta: 8.3h
iter: 750, acc (bg: 1.00, fg: 0.90, iou: 0.81), loss (bbox_3d: 0.7230, cls: 0.1290, iou: 0.2170), misc (ry: 0.73, z: 1.14), dt: 0.60, eta: 8.2h
iter: 1000, acc (bg: 1.00, fg: 0.92, iou: 0.83), loss (bbox_3d: 0.5881, cls: 0.1109, iou: 0.1957), misc (ry: 0.62, z: 1.00), dt: 0.60, eta: 8.2h
iter: 1250, acc (bg: 1.00, fg: 0.91, iou: 0.83), loss (bbox_3d: 0.5223, cls: 0.1170, iou: 0.1938), misc (ry: 0.60, z: 0.85), dt: 0.61, eta: 8.2h
iter: 1500, acc (bg: 1.00, fg: 0.93, iou: 0.84), loss (bbox_3d: 0.5032, cls: 0.0980, iou: 0.1865), misc (ry: 0.54, z: 0.85), dt: 0.61, eta: 8.2h
iter: 1750, acc (bg: 1.00, fg: 0.93, iou: 0.84), loss (bbox_3d: 0.4491, cls: 0.0911, iou: 0.1781), misc (ry: 0.50, z: 0.86), dt: 0.61, eta: 8.1h
iter: 2000, acc (bg: 1.00, fg: 0.95, iou: 0.85), loss (bbox_3d: 0.4504, cls: 0.0807, iou: 0.1697), misc (ry: 0.53, z: 0.79), dt: 0.61, eta: 8.1h
iter: 2250, acc (bg: 1.00, fg: 0.95, iou: 0.85), loss (bbox_3d: 0.4343, cls: 0.0779, iou: 0.1673), misc (ry: 0.57, z: 0.72), dt: 0.61, eta: 8.0h
iter: 2500, acc (bg: 1.00, fg: 0.95, iou: 0.86), loss (bbox_3d: 0.3655, cls: 0.0754, iou: 0.1592), misc (ry: 0.49, z: 0.68), dt: 0.61, eta: 8.0h
iter: 2750, acc (bg: 1.00, fg: 0.96, iou: 0.87), loss (bbox_3d: 0.3271, cls: 0.0675, iou: 0.1484), misc (ry: 0.41, z: 0.69), dt: 0.61, eta: 8.0h
iter: 3000, acc (bg: 1.00, fg: 0.96, iou: 0.87), loss (bbox_3d: 0.2997, cls: 0.0649, iou: 0.1461), misc (ry: 0.43, z: 0.64), dt: 0.61, eta: 7.9h
iter: 3250, acc (bg: 1.00, fg: 0.95, iou: 0.86), loss (bbox_3d: 0.3162, cls: 0.0725, iou: 0.1526), misc (ry: 0.42, z: 0.67), dt: 0.61, eta: 7.9h
iter: 3500, acc (bg: 1.00, fg: 0.96, iou: 0.87), loss (bbox_3d: 0.2788, cls: 0.0619, iou: 0.1438), misc (ry: 0.34, z: 0.63), dt: 0.61, eta: 7.8h
iter: 3750, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2570, cls: 0.0607, iou: 0.1422), misc (ry: 0.32, z: 0.67), dt: 0.61, eta: 7.8h
iter: 4000, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2221, cls: 0.0536, iou: 0.1324), misc (ry: 0.31, z: 0.57), dt: 0.61, eta: 7.7h
iter: 4250, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2086, cls: 0.0543, iou: 0.1339), misc (ry: 0.31, z: 0.57), dt: 0.61, eta: 7.7h
iter: 4500, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2446, cls: 0.0577, iou: 0.1327), misc (ry: 0.37, z: 0.54), dt: 0.61, eta: 7.7h
iter: 4750, acc (bg: 1.00, fg: 0.96, iou: 0.87), loss (bbox_3d: 0.3091, cls: 0.0629, iou: 0.1483), misc (ry: 0.49, z: 0.62), dt: 0.61, eta: 7.6h
iter: 5000, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2930, cls: 0.0616, iou: 0.1422), misc (ry: 0.40, z: 0.65), dt: 0.61, eta: 7.6h
iter: 5250, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2990, cls: 0.0585, iou: 0.1399), misc (ry: 0.46, z: 0.65), dt: 0.61, eta: 7.6h
iter: 5500, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2527, cls: 0.0610, iou: 0.1368), misc (ry: 0.35, z: 0.62), dt: 0.61, eta: 7.5h
iter: 5750, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2377, cls: 0.0587, iou: 0.1353), misc (ry: 0.29, z: 0.62), dt: 0.61, eta: 7.5h
iter: 6000, acc (bg: 1.00, fg: 0.97, iou: 0.87), loss (bbox_3d: 0.2737, cls: 0.0580, iou: 0.1414), misc (ry: 0.36, z: 0.65), dt: 0.61, eta: 7.4h
iter: 6250, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2346, cls: 0.0552, iou: 0.1322), misc (ry: 0.31, z: 0.60), dt: 0.61, eta: 7.4h
iter: 6500, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.1966, cls: 0.0533, iou: 0.1274), misc (ry: 0.30, z: 0.52), dt: 0.61, eta: 7.4h
iter: 6750, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2216, cls: 0.0542, iou: 0.1291), misc (ry: 0.33, z: 0.56), dt: 0.61, eta: 7.3h
iter: 7000, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2290, cls: 0.0561, iou: 0.1317), misc (ry: 0.35, z: 0.54), dt: 0.61, eta: 7.3h
iter: 7250, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1932, cls: 0.0497, iou: 0.1247), misc (ry: 0.31, z: 0.52), dt: 0.61, eta: 7.2h
iter: 7500, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1808, cls: 0.0499, iou: 0.1212), misc (ry: 0.29, z: 0.49), dt: 0.61, eta: 7.2h
iter: 7750, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1867, cls: 0.0499, iou: 0.1213), misc (ry: 0.30, z: 0.48), dt: 0.61, eta: 7.1h
iter: 8000, acc (bg: 1.00, fg: 0.97, iou: 0.89), loss (bbox_3d: 0.1908, cls: 0.0534, iou: 0.1252), misc (ry: 0.29, z: 0.53), dt: 0.61, eta: 7.1h
iter: 8250, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1811, cls: 0.0489, iou: 0.1207), misc (ry: 0.28, z: 0.49), dt: 0.61, eta: 7.1h
iter: 8500, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1719, cls: 0.0455, iou: 0.1203), misc (ry: 0.27, z: 0.55), dt: 0.61, eta: 7.0h
iter: 8750, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1664, cls: 0.0453, iou: 0.1158), misc (ry: 0.27, z: 0.49), dt: 0.61, eta: 7.0h
iter: 9000, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1573, cls: 0.0452, iou: 0.1167), misc (ry: 0.24, z: 0.51), dt: 0.61, eta: 6.9h
iter: 9250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1583, cls: 0.0435, iou: 0.1134), misc (ry: 0.24, z: 0.52), dt: 0.61, eta: 6.9h
iter: 9500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1520, cls: 0.0463, iou: 0.1132), misc (ry: 0.22, z: 0.49), dt: 0.61, eta: 6.8h
iter: 9750, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1550, cls: 0.0469, iou: 0.1159), misc (ry: 0.25, z: 0.47), dt: 0.61, eta: 6.8h
iter: 10000, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1492, cls: 0.0457, iou: 0.1125), misc (ry: 0.23, z: 0.46), dt: 0.61, eta: 6.8h
testing 1000/3769, dt: 0.177, eta: 8.2m
testing 2000/3769, dt: 0.177, eta: 5.2m
testing 3000/3769, dt: 0.177, eta: 2.3m
test_iter 10000 2d car --> easy: 0.1475, mod: 0.1422, hard: 0.1290
test_iter 10000 gr car --> easy: 0.1172, mod: 0.1102, hard: 0.1047
test_iter 10000 3d car --> easy: 0.1107, mod: 0.1045, hard: 0.1000
test_iter 10000 2d pedestrian --> easy: 0.1319, mod: 0.1275, hard: 0.1191
test_iter 10000 gr pedestrian --> easy: 0.0338, mod: 0.0303, hard: 0.0287
test_iter 10000 3d pedestrian --> easy: 0.0263, mod: 0.0246, hard: 0.0216
test_iter 10000 2d cyclist --> easy: 0.0070, mod: 0.0933, hard: 0.0934
test_iter 10000 gr cyclist --> easy: 0.0005, mod: 0.0091, hard: 0.0091
test_iter 10000 3d cyclist --> easy: 0.0004, mod: 0.0091, hard: 0.0091
iter: 10250, acc (bg: 1.00, fg: 0.97, iou: 0.89), loss (bbox_3d: 0.1634, cls: 0.0523, iou: 0.1151), misc (ry: 0.23, z: 0.47), dt: 0.68, eta: 7.5h
iter: 10500, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1461, cls: 0.0464, iou: 0.1145), misc (ry: 0.22, z: 0.48), dt: 0.68, eta: 7.5h
iter: 10750, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1367, cls: 0.0417, iou: 0.1102), misc (ry: 0.24, z: 0.45), dt: 0.68, eta: 7.4h
iter: 11000, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1285, cls: 0.0425, iou: 0.1105), misc (ry: 0.21, z: 0.45), dt: 0.68, eta: 7.4h
iter: 11250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1389, cls: 0.0466, iou: 0.1107), misc (ry: 0.24, z: 0.42), dt: 0.68, eta: 7.3h
iter: 11500, acc (bg: 1.00, fg: 0.97, iou: 0.89), loss (bbox_3d: 0.1779, cls: 0.0499, iou: 0.1184), misc (ry: 0.29, z: 0.47), dt: 0.68, eta: 7.2h
iter: 11750, acc (bg: 1.00, fg: 0.97, iou: 0.88), loss (bbox_3d: 0.2296, cls: 0.0552, iou: 0.1349), misc (ry: 0.35, z: 0.54), dt: 0.67, eta: 7.2h
iter: 12000, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1879, cls: 0.0477, iou: 0.1214), misc (ry: 0.31, z: 0.55), dt: 0.67, eta: 7.1h
iter: 12250, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1771, cls: 0.0469, iou: 0.1202), misc (ry: 0.29, z: 0.50), dt: 0.67, eta: 7.0h
iter: 12500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1483, cls: 0.0453, iou: 0.1120), misc (ry: 0.25, z: 0.45), dt: 0.67, eta: 7.0h
iter: 12750, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1304, cls: 0.0440, iou: 0.1066), misc (ry: 0.21, z: 0.45), dt: 0.67, eta: 6.9h
iter: 13000, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1387, cls: 0.0433, iou: 0.1078), misc (ry: 0.25, z: 0.46), dt: 0.67, eta: 6.9h
iter: 13250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1455, cls: 0.0426, iou: 0.1072), misc (ry: 0.25, z: 0.45), dt: 0.67, eta: 6.8h
iter: 13500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1302, cls: 0.0426, iou: 0.1091), misc (ry: 0.23, z: 0.45), dt: 0.67, eta: 6.7h
iter: 13750, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1273, cls: 0.0419, iou: 0.1045), misc (ry: 0.21, z: 0.46), dt: 0.66, eta: 6.7h
iter: 14000, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1262, cls: 0.0423, iou: 0.1062), misc (ry: 0.21, z: 0.44), dt: 0.66, eta: 6.6h
iter: 14250, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1056, cls: 0.0402, iou: 0.0967), misc (ry: 0.19, z: 0.41), dt: 0.66, eta: 6.6h
iter: 14500, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1046, cls: 0.0400, iou: 0.0982), misc (ry: 0.19, z: 0.39), dt: 0.66, eta: 6.5h
iter: 14750, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1010, cls: 0.0400, iou: 0.0969), misc (ry: 0.19, z: 0.37), dt: 0.66, eta: 6.5h
iter: 15000, acc (bg: 1.00, fg: 0.97, iou: 0.89), loss (bbox_3d: 0.2119, cls: 0.0535, iou: 0.1205), misc (ry: 0.36, z: 0.53), dt: 0.66, eta: 6.4h
iter: 15250, acc (bg: 1.00, fg: 0.98, iou: 0.89), loss (bbox_3d: 0.1946, cls: 0.0482, iou: 0.1169), misc (ry: 0.34, z: 0.50), dt: 0.66, eta: 6.4h
iter: 15500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1533, cls: 0.0434, iou: 0.1103), misc (ry: 0.24, z: 0.48), dt: 0.66, eta: 6.3h
iter: 15750, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1477, cls: 0.0452, iou: 0.1092), misc (ry: 0.26, z: 0.47), dt: 0.66, eta: 6.3h
iter: 16000, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1592, cls: 0.0460, iou: 0.1135), misc (ry: 0.27, z: 0.48), dt: 0.66, eta: 6.2h
iter: 16250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1421, cls: 0.0429, iou: 0.1058), misc (ry: 0.24, z: 0.47), dt: 0.66, eta: 6.2h
iter: 16500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1281, cls: 0.0433, iou: 0.1054), misc (ry: 0.21, z: 0.46), dt: 0.66, eta: 6.1h
iter: 16750, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.1164, cls: 0.0371, iou: 0.1007), misc (ry: 0.20, z: 0.42), dt: 0.66, eta: 6.1h
iter: 17000, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1199, cls: 0.0415, iou: 0.1016), misc (ry: 0.21, z: 0.42), dt: 0.65, eta: 6.0h
iter: 17250, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1317, cls: 0.0403, iou: 0.1029), misc (ry: 0.24, z: 0.45), dt: 0.65, eta: 5.9h
iter: 17500, acc (bg: 1.00, fg: 0.98, iou: 0.90), loss (bbox_3d: 0.1198, cls: 0.0432, iou: 0.1042), misc (ry: 0.20, z: 0.40), dt: 0.65, eta: 5.9h
iter: 17750, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1228, cls: 0.0401, iou: 0.1009), misc (ry: 0.22, z: 0.42), dt: 0.65, eta: 5.8h
iter: 18000, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1024, cls: 0.0395, iou: 0.0976), misc (ry: 0.18, z: 0.43), dt: 0.65, eta: 5.8h
iter: 18250, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1087, cls: 0.0384, iou: 0.0941), misc (ry: 0.18, z: 0.44), dt: 0.65, eta: 5.7h
iter: 18500, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.1012, cls: 0.0381, iou: 0.0939), misc (ry: 0.19, z: 0.39), dt: 0.65, eta: 5.7h
iter: 18750, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.0900, cls: 0.0362, iou: 0.0929), misc (ry: 0.18, z: 0.37), dt: 0.65, eta: 5.6h
iter: 19000, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.1004, cls: 0.0382, iou: 0.0958), misc (ry: 0.19, z: 0.38), dt: 0.65, eta: 5.6h
iter: 19250, acc (bg: 1.00, fg: 0.98, iou: 0.91), loss (bbox_3d: 0.1298, cls: 0.0388, iou: 0.1000), misc (ry: 0.24, z: 0.42), dt: 0.65, eta: 5.5h
iter: 19500, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.1086, cls: 0.0383, iou: 0.0963), misc (ry: 0.21, z: 0.40), dt: 0.65, eta: 5.5h
iter: 19750, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.1037, cls: 0.0384, iou: 0.0925), misc (ry: 0.21, z: 0.35), dt: 0.65, eta: 5.5h
iter: 20000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0911, cls: 0.0377, iou: 0.0907), misc (ry: 0.19, z: 0.35), dt: 0.65, eta: 5.4h
testing 1000/3769, dt: 0.177, eta: 8.2m
testing 2000/3769, dt: 0.177, eta: 5.2m
testing 3000/3769, dt: 0.177, eta: 2.3m
test_iter 20000 2d car --> easy: 0.1021, mod: 0.0926, hard: 0.0819
test_iter 20000 gr car --> easy: 0.0704, mod: 0.0619, hard: 0.0562
test_iter 20000 3d car --> easy: 0.0651, mod: 0.0579, hard: 0.0539
test_iter 20000 2d pedestrian --> easy: 0.1326, mod: 0.1261, hard: 0.1182
test_iter 20000 gr pedestrian --> easy: 0.0577, mod: 0.0545, hard: 0.0522
test_iter 20000 3d pedestrian --> easy: 0.0560, mod: 0.0533, hard: 0.0515
test_iter 20000 2d cyclist --> easy: 0.1067, mod: 0.0972, hard: 0.0958
test_iter 20000 gr cyclist --> easy: 0.0069, mod: 0.0038, hard: 0.0039
test_iter 20000 3d cyclist --> easy: 0.0065, mod: 0.0038, hard: 0.0039
iter: 20250, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.1049, cls: 0.0387, iou: 0.0936), misc (ry: 0.23, z: 0.38), dt: 0.69, eta: 5.7h
iter: 20500, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.0923, cls: 0.0365, iou: 0.0941), misc (ry: 0.17, z: 0.36), dt: 0.68, eta: 5.6h
iter: 20750, acc (bg: 1.00, fg: 0.99, iou: 0.91), loss (bbox_3d: 0.0919, cls: 0.0374, iou: 0.0922), misc (ry: 0.19, z: 0.35), dt: 0.68, eta: 5.6h
iter: 21000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0811, cls: 0.0366, iou: 0.0887), misc (ry: 0.16, z: 0.34), dt: 0.68, eta: 5.5h
iter: 21250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0823, cls: 0.0364, iou: 0.0894), misc (ry: 0.16, z: 0.34), dt: 0.68, eta: 5.4h
iter: 21500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0814, cls: 0.0357, iou: 0.0864), misc (ry: 0.16, z: 0.34), dt: 0.68, eta: 5.4h
iter: 21750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0846, cls: 0.0362, iou: 0.0894), misc (ry: 0.16, z: 0.36), dt: 0.68, eta: 5.3h
iter: 22000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0829, cls: 0.0353, iou: 0.0864), misc (ry: 0.16, z: 0.34), dt: 0.68, eta: 5.3h
iter: 22250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0761, cls: 0.0346, iou: 0.0853), misc (ry: 0.16, z: 0.32), dt: 0.68, eta: 5.2h
iter: 22500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0781, cls: 0.0357, iou: 0.0863), misc (ry: 0.16, z: 0.36), dt: 0.68, eta: 5.2h
iter: 22750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0810, cls: 0.0378, iou: 0.0866), misc (ry: 0.16, z: 0.35), dt: 0.68, eta: 5.1h
iter: 23000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0839, cls: 0.0357, iou: 0.0892), misc (ry: 0.16, z: 0.37), dt: 0.68, eta: 5.1h
iter: 23250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0779, cls: 0.0368, iou: 0.0839), misc (ry: 0.15, z: 0.34), dt: 0.68, eta: 5.0h
iter: 23500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0822, cls: 0.0358, iou: 0.0875), misc (ry: 0.16, z: 0.38), dt: 0.68, eta: 5.0h
iter: 23750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0739, cls: 0.0346, iou: 0.0829), misc (ry: 0.16, z: 0.35), dt: 0.67, eta: 4.9h
iter: 24000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0681, cls: 0.0315, iou: 0.0809), misc (ry: 0.14, z: 0.34), dt: 0.67, eta: 4.9h
iter: 24250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0731, cls: 0.0364, iou: 0.0846), misc (ry: 0.15, z: 0.31), dt: 0.67, eta: 4.8h
iter: 24500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0655, cls: 0.0329, iou: 0.0777), misc (ry: 0.14, z: 0.33), dt: 0.67, eta: 4.8h
iter: 24750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0740, cls: 0.0355, iou: 0.0843), misc (ry: 0.15, z: 0.31), dt: 0.67, eta: 4.7h
iter: 25000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0612, cls: 0.0338, iou: 0.0799), misc (ry: 0.14, z: 0.29), dt: 0.67, eta: 4.7h
iter: 25250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0705, cls: 0.0334, iou: 0.0807), misc (ry: 0.14, z: 0.32), dt: 0.67, eta: 4.6h
iter: 25500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0728, cls: 0.0335, iou: 0.0828), misc (ry: 0.16, z: 0.32), dt: 0.67, eta: 4.6h
iter: 25750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0821, cls: 0.0343, iou: 0.0834), misc (ry: 0.17, z: 0.34), dt: 0.67, eta: 4.5h
iter: 26000, acc (bg: 1.00, fg: 0.98, iou: 0.92), loss (bbox_3d: 0.1027, cls: 0.0410, iou: 0.0906), misc (ry: 0.19, z: 0.38), dt: 0.67, eta: 4.5h
iter: 26250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0879, cls: 0.0360, iou: 0.0894), misc (ry: 0.17, z: 0.34), dt: 0.67, eta: 4.4h
iter: 26500, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0791, cls: 0.0360, iou: 0.0843), misc (ry: 0.16, z: 0.32), dt: 0.67, eta: 4.4h
iter: 26750, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0748, cls: 0.0358, iou: 0.0845), misc (ry: 0.16, z: 0.32), dt: 0.67, eta: 4.3h
iter: 27000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0729, cls: 0.0347, iou: 0.0813), misc (ry: 0.15, z: 0.33), dt: 0.67, eta: 4.3h
iter: 27250, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0727, cls: 0.0323, iou: 0.0831), misc (ry: 0.14, z: 0.36), dt: 0.67, eta: 4.2h
iter: 27500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0608, cls: 0.0313, iou: 0.0783), misc (ry: 0.13, z: 0.33), dt: 0.67, eta: 4.2h
iter: 27750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0589, cls: 0.0316, iou: 0.0769), misc (ry: 0.13, z: 0.30), dt: 0.67, eta: 4.1h
iter: 28000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0628, cls: 0.0333, iou: 0.0793), misc (ry: 0.14, z: 0.31), dt: 0.67, eta: 4.1h
iter: 28250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0622, cls: 0.0334, iou: 0.0779), misc (ry: 0.15, z: 0.30), dt: 0.66, eta: 4.0h
iter: 28500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0578, cls: 0.0330, iou: 0.0760), misc (ry: 0.13, z: 0.30), dt: 0.66, eta: 4.0h
iter: 28750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0648, cls: 0.0336, iou: 0.0767), misc (ry: 0.15, z: 0.30), dt: 0.66, eta: 3.9h
iter: 29000, acc (bg: 1.00, fg: 0.99, iou: 0.92), loss (bbox_3d: 0.0683, cls: 0.0341, iou: 0.0810), misc (ry: 0.16, z: 0.32), dt: 0.66, eta: 3.9h
iter: 29250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0740, cls: 0.0348, iou: 0.0792), misc (ry: 0.17, z: 0.32), dt: 0.66, eta: 3.8h
iter: 29500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0675, cls: 0.0340, iou: 0.0781), misc (ry: 0.15, z: 0.29), dt: 0.66, eta: 3.8h
iter: 29750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0671, cls: 0.0335, iou: 0.0796), misc (ry: 0.15, z: 0.30), dt: 0.66, eta: 3.7h
iter: 30000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0611, cls: 0.0326, iou: 0.0765), misc (ry: 0.14, z: 0.30), dt: 0.66, eta: 3.7h
testing 1000/3769, dt: 0.176, eta: 8.1m
testing 2000/3769, dt: 0.176, eta: 5.2m
testing 3000/3769, dt: 0.176, eta: 2.3m
test_iter 30000 2d car --> easy: 0.1104, mod: 0.1051, hard: 0.0953
test_iter 30000 gr car --> easy: 0.0826, mod: 0.0758, hard: 0.0709
test_iter 30000 3d car --> easy: 0.0769, mod: 0.0723, hard: 0.0688
test_iter 30000 2d pedestrian --> easy: 0.1360, mod: 0.1335, hard: 0.1230
test_iter 30000 gr pedestrian --> easy: 0.0573, mod: 0.0532, hard: 0.0522
test_iter 30000 3d pedestrian --> easy: 0.0553, mod: 0.0524, hard: 0.0514
test_iter 30000 2d cyclist --> easy: 0.1094, mod: 0.0989, hard: 0.0982
test_iter 30000 gr cyclist --> easy: 0.0261, mod: 0.0245, hard: 0.0227
test_iter 30000 3d cyclist --> easy: 0.0261, mod: 0.0244, hard: 0.0227
iter: 30250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0574, cls: 0.0320, iou: 0.0738), misc (ry: 0.14, z: 0.29), dt: 0.69, eta: 3.8h
iter: 30500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0582, cls: 0.0332, iou: 0.0733), misc (ry: 0.14, z: 0.29), dt: 0.69, eta: 3.7h
iter: 30750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0555, cls: 0.0315, iou: 0.0735), misc (ry: 0.13, z: 0.29), dt: 0.69, eta: 3.7h
iter: 31000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0548, cls: 0.0305, iou: 0.0751), misc (ry: 0.13, z: 0.30), dt: 0.69, eta: 3.6h
iter: 31250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0563, cls: 0.0325, iou: 0.0739), misc (ry: 0.12, z: 0.29), dt: 0.68, eta: 3.6h
iter: 31500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0675, cls: 0.0349, iou: 0.0781), misc (ry: 0.17, z: 0.32), dt: 0.68, eta: 3.5h
iter: 31750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0521, cls: 0.0304, iou: 0.0704), misc (ry: 0.14, z: 0.28), dt: 0.68, eta: 3.5h
iter: 32000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0557, cls: 0.0309, iou: 0.0732), misc (ry: 0.14, z: 0.29), dt: 0.68, eta: 3.4h
iter: 32250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0525, cls: 0.0315, iou: 0.0705), misc (ry: 0.13, z: 0.28), dt: 0.68, eta: 3.4h
iter: 32500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0536, cls: 0.0305, iou: 0.0716), misc (ry: 0.14, z: 0.28), dt: 0.68, eta: 3.3h
iter: 32750, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0521, cls: 0.0314, iou: 0.0715), misc (ry: 0.13, z: 0.28), dt: 0.68, eta: 3.3h
iter: 33000, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0494, cls: 0.0295, iou: 0.0700), misc (ry: 0.13, z: 0.27), dt: 0.68, eta: 3.2h
iter: 33250, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0503, cls: 0.0301, iou: 0.0687), misc (ry: 0.12, z: 0.27), dt: 0.68, eta: 3.2h
iter: 33500, acc (bg: 1.00, fg: 0.99, iou: 0.93), loss (bbox_3d: 0.0518, cls: 0.0302, iou: 0.0702), misc (ry: 0.12, z: 0.28), dt: 0.68, eta: 3.1h
iter: 33750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0474, cls: 0.0300, iou: 0.0661), misc (ry: 0.13, z: 0.26), dt: 0.68, eta: 3.1h
iter: 34000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0518, cls: 0.0309, iou: 0.0686), misc (ry: 0.13, z: 0.27), dt: 0.68, eta: 3.0h
iter: 34250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0464, cls: 0.0310, iou: 0.0682), misc (ry: 0.12, z: 0.26), dt: 0.68, eta: 3.0h
iter: 34500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0489, cls: 0.0314, iou: 0.0671), misc (ry: 0.12, z: 0.26), dt: 0.68, eta: 2.9h
iter: 34750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0482, cls: 0.0289, iou: 0.0678), misc (ry: 0.12, z: 0.25), dt: 0.68, eta: 2.9h
iter: 35000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0453, cls: 0.0313, iou: 0.0675), misc (ry: 0.12, z: 0.24), dt: 0.68, eta: 2.8h
iter: 35250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0453, cls: 0.0315, iou: 0.0669), misc (ry: 0.12, z: 0.25), dt: 0.68, eta: 2.8h
iter: 35500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0493, cls: 0.0319, iou: 0.0670), misc (ry: 0.12, z: 0.28), dt: 0.68, eta: 2.7h
iter: 35750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0458, cls: 0.0296, iou: 0.0662), misc (ry: 0.12, z: 0.27), dt: 0.68, eta: 2.7h
iter: 36000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0467, cls: 0.0290, iou: 0.0671), misc (ry: 0.12, z: 0.26), dt: 0.68, eta: 2.6h
iter: 36250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0385, cls: 0.0289, iou: 0.0617), misc (ry: 0.11, z: 0.24), dt: 0.67, eta: 2.6h
iter: 36500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0457, cls: 0.0293, iou: 0.0652), misc (ry: 0.11, z: 0.26), dt: 0.67, eta: 2.5h
iter: 36750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0470, cls: 0.0308, iou: 0.0668), misc (ry: 0.12, z: 0.24), dt: 0.67, eta: 2.5h
iter: 37000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0409, cls: 0.0293, iou: 0.0625), misc (ry: 0.11, z: 0.24), dt: 0.67, eta: 2.4h
iter: 37250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0508, cls: 0.0309, iou: 0.0640), misc (ry: 0.13, z: 0.26), dt: 0.67, eta: 2.4h
iter: 37500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0442, cls: 0.0309, iou: 0.0650), misc (ry: 0.12, z: 0.24), dt: 0.67, eta: 2.3h
iter: 37750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0411, cls: 0.0287, iou: 0.0626), misc (ry: 0.11, z: 0.26), dt: 0.67, eta: 2.3h
iter: 38000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0410, cls: 0.0293, iou: 0.0624), misc (ry: 0.11, z: 0.23), dt: 0.67, eta: 2.2h
iter: 38250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0378, cls: 0.0294, iou: 0.0601), misc (ry: 0.11, z: 0.24), dt: 0.67, eta: 2.2h
iter: 38500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0379, cls: 0.0280, iou: 0.0600), misc (ry: 0.11, z: 0.23), dt: 0.67, eta: 2.1h
iter: 38750, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0445, cls: 0.0317, iou: 0.0623), misc (ry: 0.12, z: 0.24), dt: 0.67, eta: 2.1h
iter: 39000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0390, cls: 0.0298, iou: 0.0603), misc (ry: 0.11, z: 0.23), dt: 0.67, eta: 2.0h
iter: 39250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0400, cls: 0.0295, iou: 0.0613), misc (ry: 0.11, z: 0.24), dt: 0.67, eta: 2.0h
iter: 39500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0381, cls: 0.0300, iou: 0.0586), misc (ry: 0.11, z: 0.22), dt: 0.67, eta: 2.0h
iter: 39750, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0360, cls: 0.0269, iou: 0.0571), misc (ry: 0.10, z: 0.22), dt: 0.67, eta: 1.9h
iter: 40000, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0358, cls: 0.0283, iou: 0.0580), misc (ry: 0.10, z: 0.23), dt: 0.67, eta: 1.9h
testing 1000/3769, dt: 0.174, eta: 8.0m
testing 2000/3769, dt: 0.174, eta: 5.1m
testing 3000/3769, dt: 0.173, eta: 2.2m
test_iter 40000 2d car --> easy: 0.1404, mod: 0.1399, hard: 0.1264
test_iter 40000 gr car --> easy: 0.1139, mod: 0.1073, hard: 0.1021
test_iter 40000 3d car --> easy: 0.1089, mod: 0.1037, hard: 0.1000
test_iter 40000 2d pedestrian --> easy: 0.1375, mod: 0.1313, hard: 0.1235
test_iter 40000 gr pedestrian --> easy: 0.0429, mod: 0.0386, hard: 0.0369
test_iter 40000 3d pedestrian --> easy: 0.0420, mod: 0.0379, hard: 0.0342
test_iter 40000 2d cyclist --> easy: 0.0437, mod: 0.1035, hard: 0.1032
test_iter 40000 gr cyclist --> easy: 0.0194, mod: 0.0168, hard: 0.0168
test_iter 40000 3d cyclist --> easy: 0.0194, mod: 0.0168, hard: 0.0168
iter: 40250, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0415, cls: 0.0320, iou: 0.0605), misc (ry: 0.12, z: 0.22), dt: 0.69, eta: 1.9h
iter: 40500, acc (bg: 1.00, fg: 0.99, iou: 0.94), loss (bbox_3d: 0.0379, cls: 0.0302, iou: 0.0580), misc (ry: 0.11, z: 0.22), dt: 0.69, eta: 1.8h
iter: 40750, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0348, cls: 0.0269, iou: 0.0559), misc (ry: 0.10, z: 0.22), dt: 0.69, eta: 1.8h
iter: 41000, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0390, cls: 0.0292, iou: 0.0575), misc (ry: 0.11, z: 0.22), dt: 0.69, eta: 1.7h
iter: 41250, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0361, cls: 0.0302, iou: 0.0566), misc (ry: 0.11, z: 0.21), dt: 0.69, eta: 1.7h
iter: 41500, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0326, cls: 0.0282, iou: 0.0545), misc (ry: 0.10, z: 0.21), dt: 0.68, eta: 1.6h
iter: 41750, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0332, cls: 0.0271, iou: 0.0541), misc (ry: 0.10, z: 0.23), dt: 0.68, eta: 1.6h
iter: 42000, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0398, cls: 0.0294, iou: 0.0565), misc (ry: 0.10, z: 0.23), dt: 0.68, eta: 1.5h
iter: 42250, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0342, cls: 0.0284, iou: 0.0551), misc (ry: 0.10, z: 0.21), dt: 0.68, eta: 1.5h
iter: 42500, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0359, cls: 0.0274, iou: 0.0556), misc (ry: 0.10, z: 0.21), dt: 0.68, eta: 1.4h
iter: 42750, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0382, cls: 0.0305, iou: 0.0570), misc (ry: 0.11, z: 0.21), dt: 0.68, eta: 1.4h
iter: 43000, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0322, cls: 0.0273, iou: 0.0517), misc (ry: 0.10, z: 0.22), dt: 0.68, eta: 1.3h
iter: 43250, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0307, cls: 0.0277, iou: 0.0515), misc (ry: 0.10, z: 0.20), dt: 0.68, eta: 1.3h
iter: 43500, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0324, cls: 0.0296, iou: 0.0534), misc (ry: 0.10, z: 0.20), dt: 0.68, eta: 1.2h
iter: 43750, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0309, cls: 0.0277, iou: 0.0512), misc (ry: 0.10, z: 0.19), dt: 0.68, eta: 1.2h
iter: 44000, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0336, cls: 0.0287, iou: 0.0514), misc (ry: 0.10, z: 0.21), dt: 0.68, eta: 1.1h
iter: 44250, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0326, cls: 0.0277, iou: 0.0520), misc (ry: 0.09, z: 0.21), dt: 0.68, eta: 1.1h
iter: 44500, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0299, cls: 0.0264, iou: 0.0507), misc (ry: 0.10, z: 0.20), dt: 0.68, eta: 1.0h
iter: 44750, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0303, cls: 0.0278, iou: 0.0495), misc (ry: 0.10, z: 0.19), dt: 0.68, eta: 59.5m
iter: 45000, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0299, cls: 0.0269, iou: 0.0487), misc (ry: 0.10, z: 0.20), dt: 0.68, eta: 56.6m
iter: 45250, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0325, cls: 0.0286, iou: 0.0506), misc (ry: 0.10, z: 0.20), dt: 0.68, eta: 53.8m
iter: 45500, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0308, cls: 0.0284, iou: 0.0494), misc (ry: 0.09, z: 0.20), dt: 0.68, eta: 50.9m
iter: 45750, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0299, cls: 0.0276, iou: 0.0483), misc (ry: 0.09, z: 0.20), dt: 0.68, eta: 48.1m
iter: 46000, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0317, cls: 0.0267, iou: 0.0492), misc (ry: 0.10, z: 0.20), dt: 0.68, eta: 45.2m
iter: 46250, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0317, cls: 0.0283, iou: 0.0490), misc (ry: 0.09, z: 0.20), dt: 0.68, eta: 42.4m
iter: 46500, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0316, cls: 0.0282, iou: 0.0494), misc (ry: 0.10, z: 0.20), dt: 0.68, eta: 39.5m
iter: 46750, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0283, cls: 0.0263, iou: 0.0477), misc (ry: 0.09, z: 0.20), dt: 0.68, eta: 36.7m
iter: 47000, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0325, cls: 0.0279, iou: 0.0498), misc (ry: 0.10, z: 0.20), dt: 0.68, eta: 33.8m
iter: 47250, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0299, cls: 0.0276, iou: 0.0475), misc (ry: 0.10, z: 0.20), dt: 0.68, eta: 31.0m
iter: 47500, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0302, cls: 0.0278, iou: 0.0474), misc (ry: 0.09, z: 0.19), dt: 0.68, eta: 28.2m
iter: 47750, acc (bg: 1.00, fg: 0.99, iou: 0.96), loss (bbox_3d: 0.0284, cls: 0.0271, iou: 0.0466), misc (ry: 0.09, z: 0.19), dt: 0.68, eta: 25.3m
iter: 48000, acc (bg: 1.00, fg: 0.99, iou: 0.96), loss (bbox_3d: 0.0292, cls: 0.0271, iou: 0.0468), misc (ry: 0.09, z: 0.19), dt: 0.68, eta: 22.5m
iter: 48250, acc (bg: 1.00, fg: 0.99, iou: 0.95), loss (bbox_3d: 0.0305, cls: 0.0288, iou: 0.0478), misc (ry: 0.09, z: 0.19), dt: 0.68, eta: 19.7m
iter: 48500, acc (bg: 1.00, fg: 0.99, iou: 0.96), loss (bbox_3d: 0.0286, cls: 0.0261, iou: 0.0452), misc (ry: 0.09, z: 0.19), dt: 0.67, eta: 16.9m
iter: 48750, acc (bg: 1.00, fg: 0.99, iou: 0.96), loss (bbox_3d: 0.0298, cls: 0.0266, iou: 0.0467), misc (ry: 0.09, z: 0.19), dt: 0.67, eta: 14.1m
iter: 49000, acc (bg: 1.00, fg: 0.99, iou: 0.96), loss (bbox_3d: 0.0267, cls: 0.0272, iou: 0.0437), misc (ry: 0.09, z: 0.19), dt: 0.67, eta: 11.2m
iter: 49250, acc (bg: 1.00, fg: 0.99, iou: 0.96), loss (bbox_3d: 0.0272, cls: 0.0269, iou: 0.0447), misc (ry: 0.09, z: 0.18), dt: 0.67, eta: 8.4m
iter: 49500, acc (bg: 1.00, fg: 0.99, iou: 0.96), loss (bbox_3d: 0.0271, cls: 0.0270, iou: 0.0446), misc (ry: 0.09, z: 0.18), dt: 0.67, eta: 5.6m
iter: 49750, acc (bg: 1.00, fg: 0.99, iou: 0.96), loss (bbox_3d: 0.0285, cls: 0.0269, iou: 0.0447), misc (ry: 0.09, z: 0.19), dt: 0.67, eta: 2.8m
iter: 50000, acc (bg: 1.00, fg: 0.99, iou: 0.96), loss (bbox_3d: 0.0261, cls: 0.0262, iou: 0.0428), misc (ry: 0.09, z: 0.17), dt: 0.67, eta: 0.7s
testing 1000/3769, dt: 0.174, eta: 8.0m
testing 2000/3769, dt: 0.174, eta: 5.1m
testing 3000/3769, dt: 0.174, eta: 2.2m
test_iter 50000 2d car --> easy: 0.1230, mod: 0.1212, hard: 0.1068
test_iter 50000 gr car --> easy: 0.0936, mod: 0.0860, hard: 0.0806
test_iter 50000 3d car --> easy: 0.0888, mod: 0.0829, hard: 0.0783
test_iter 50000 2d pedestrian --> easy: 0.1367, mod: 0.1337, hard: 0.1254
test_iter 50000 gr pedestrian --> easy: 0.1032, mod: 0.0997, hard: 0.0984
test_iter 50000 3d pedestrian --> easy: 0.1013, mod: 0.0986, hard: 0.0951
test_iter 50000 2d cyclist --> easy: 0.0299, mod: 0.1033, hard: 0.1020
test_iter 50000 gr cyclist --> easy: 0.0090, mod: 0.0053, hard: 0.0053
test_iter 50000 3d cyclist --> easy: 0.0070, mod: 0.0050, hard: 0.0050
Segmentation fault (core dumped)
I also notice there is a segmentation fault at the bottom of output. I guess that might come from the permission I have on my side - I am a user on a big machine. If that is an error causes the low performance, please let me know.
Thanks in advance!
Best,
Shukai
Failure to reproduce results in the paper
The results on my computer is :
test_iter tmp_results 2d car --> easy: 0.9246, mod: 0.8458, hard: 0.6830
test_iter tmp_results gr car --> easy: 0.2729, mod: 0.2161, hard: 0.1795
test_iter tmp_results 3d car --> easy: 0.2052, mod: 0.1654, hard: 0.1445
test_iter tmp_results 2d pedestrian --> easy: 0.6582, mod: 0.5779, hard: 0.4972
test_iter tmp_results gr pedestrian --> easy: 0.0534, mod: 0.0558, hard: 0.0522
test_iter tmp_results 3d pedestrian --> easy: 0.0483, mod: 0.0513, hard: 0.0467
test_iter tmp_results 2d cyclist --> easy: 0.6431, mod: 0.4087, hard: 0.4072
test_iter tmp_results gr cyclist --> easy: 0.0312, mod: 0.0279, hard: 0.0292
test_iter tmp_results 3d cyclist --> easy: 0.0309, mod: 0.0267, hard: 0.0280
There is a big gap between pedestrians and cyclists in the paper, I don't know why.
No such file or directory
Thanks for the shared code!
There is a problem when I run _'sh data/kitti_split1/devkit/cpp/build.sh'
The terminator shows:
evaluate_object.cpp:12:10: fatal error: boost/numeric/ublas/matrix.hpp: No such file or directory #include <boost/numeric/ublas/matrix.hpp>
compilation terminated.
Could you please help me solve this problem?
TypeError: zip argument #1 must support iteration
Did anyone get the same problem? Or did anyone has ideas?
File "scripts/train_rpn_3d.py", line 196, in <module>
main(sys.argv[1:])
File "scripts/train_rpn_3d.py", line 122, in main
cls, prob, bbox_2d, bbox_3d, feat_size = rpn_net(images)
File "/home/xxx/miniconda3/envs/Monocular3D/lib/python3.7/site-packages/torch/nn/modules/module.py", line 550, in __call__
result = self.forward(*input, **kwargs)
File "/home/xxx/miniconda3/envs/Monocular3D/lib/python3.7/site-packages/torch/nn/parallel/data_parallel.py", line 156, in forward
return self.gather(outputs, self.output_device)
File "/home/xxx/miniconda3/envs/Monocular3D/lib/python3.7/site-packages/torch/nn/parallel/data_parallel.py", line 168, in gather
return gather(outputs, output_device, dim=self.dim)
File "/home/xxx/miniconda3/envs/Monocular3D/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 68, in gather
res = gather_map(outputs)
File "/home/xxx/miniconda3/envs/Monocular3D/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
File "/home/xxx/miniconda3/envs/Monocular3D/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
File "/home/xxx/miniconda3/envs/Monocular3D/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
TypeError: zip argument #1 must support iteration```
split files error
Hi Author, I am working on your code since last two days. in the split files i got the following error
lps@lps-Z370M-D3H:~/M3D-RPN$ python data/kitti_split1/setup_split.py
Traceback (most recent call last):
File "data/kitti_split1/setup_split.py", line 23, in
from lib.util import *
ImportError: No module named lib.util
Can you help what to import here please.
Thank you.
Error while training
Hi
I am getting this error while training. I am following the exact steps mentioned for training, i am able to perform inference but not training.
cmmd- python scripts/train_rpn_3d.py --config=kitti_3d_multi_warmup
File "scripts/train_rpn_3d.py", line 198, in
main(sys.argv[1:])
File "scripts/train_rpn_3d.py", line 124, in main
cls, prob, bbox_2d, bbox_3d, feat_size = rpn_net(images)
File "/root/utils/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/root/utils/anaconda3/lib/python3.7/site-packages/torch/nn/parallel/data_parallel.py", line 153, in forward
return self.gather(outputs, self.output_device)
File "/root/utils/anaconda3/lib/python3.7/site-packages/torch/nn/parallel/data_parallel.py", line 165, in gather
return gather(outputs, output_device, dim=self.dim)
File "/root/utils/anaconda3/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 68, in gather
res = gather_map(outputs)
File "/root/utils/anaconda3/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
File "/root/utils/anaconda3/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
File "/root/utils/anaconda3/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
TypeError: zip argument #1 must support iteration
Here is configuration of system and packaged
Ubuntu- Ubuntu 18.04.3 LTS
Cuda- 10.2
CuDNN - 7.6.5
torch - 1.4.0
python - 3.7.3
If i add os.environ["CUDA_VISIBLE_DEVICES"]="0" in training file (train_rpn_3d.py), then i don't get the above error but new error in next line. i.e.
Traceback (most recent call last):
File "scripts/train_rpn_3d.py", line 198, in
main(sys.argv[1:])
File "scripts/train_rpn_3d.py", line 127, in main
det_loss, det_stats = criterion_det(cls, prob, bbox_2d, bbox_3d, imobjs, feat_size)
File "/root/utils/anaconda3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/workspace/M3D-RPN/lib/loss/rpn_3d.py", line 125, in forward
src_anchors = self.anchors[rois[:, 4].type(torch.cuda.LongTensor), :]
File "/root/utils/anaconda3/lib/python3.7/site-packages/torch/tensor.py", line 486, in array
return self.numpy()
TypeError: can't convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.
Please let me know if this is because of some version issue or code error.
Thanks in advance.
About multi-GPU training
Hi, is your model trained with a single GPU? Have you tried using multiple GPUs?
I encountered the following error when using data-parallel in multi-GPU training.
File "scripts/train_rpn_3d.py", line 122, in main
cls, prob, bbox_2d, bbox_3d, feat_size = rpn_net(images)
File "/mnt/lustre/dingmingyu/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 493, in call
result = self.forward(*input, **kwargs)
File "/mnt/lustre/dingmingyu/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 153, in forward
return self.gather(outputs, self.output_device)
File "/mnt/lustre/dingmingyu/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 165, in gather
return gather(outputs, output_device, dim=self.dim)
File "/mnt/lustre/dingmingyu/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/scatter_gather.py", line 67, in gather
return gather_map(outputs)
File "/mnt/lustre/dingmingyu/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/scatter_gather.py", line 62, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
File "/mnt/lustre/dingmingyu/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/scatter_gather.py", line 62, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
File "/mnt/lustre/dingmingyu/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/scatter_gather.py", line 62, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
TypeError: zip argument #1 must support iteration
Do you know how to fix this? And I would like to ask about single (batch_size = 2) and multi-GPU (larger batch_size, e.g. 16) training, which one would have better performance? Thank you very much!
CUDA out of memory error
Hi Garrick,
I am trying to replicate the results. When I execute the warmup script, I get a CUDA out of memory error. I know that reducing the batch size can help in avoiding this. The batch size is 2 which is relatively low. What should be done so as to avoid this error? You can find the error that I get after executing the script below.
chinmay@chinmay-Legion-Y540-15IRH:~/Desktop/M3D-RPN$ python scripts/train_rpn_3d.py --config=kitti_3d_multi_warmup
Setting up a new session...
Visdom successfully connected to server
Preloading imdb.
weighted respectively as 1.05 and 0.00
Found 3534 foreground and 178 empty images
Labels not used in training.. ['DontCare', 'Truck', 'Tram', 'Misc', 'Person_sitting']
conf: {
model: densenet121_3d_dilate
solver_type: sgd
lr: 0.004
momentum: 0.9
weight_decay: 0.0005
max_iter: 50000
snapshot_iter: 10000
display: 250
do_test: True
lr_policy: poly
lr_steps: None
lr_target: 4e-08
rng_seed: 2
cuda_seed: 2
image_means: [0.485, 0.456, 0.406]
image_stds: [0.229, 0.224, 0.225]
feat_stride: 16
has_3d: True
test_scale: 512
crop_size: [512, 1760]
mirror_prob: 0.5
distort_prob: -1
dataset_test: kitti_split1
datasets_train: [{'anno_fmt': 'kitti_det',
'im_ext': '.png',
'name': 'kitti_split1',
'scale': 1}]
use_3d_for_2d: True
percent_anc_h: [0.0625, 0.75]
min_gt_h: 32.0
max_gt_h: 384.0
min_gt_vis: 0.65
ilbls: ['Van', 'ignore']
lbls: ['Car', 'Pedestrian', 'Cyclist']
batch_size: 2
fg_image_ratio: 1.0
box_samples: 0.2
fg_fraction: 0.2
bg_thresh_lo: 0
bg_thresh_hi: 0.5
fg_thresh: 0.5
ign_thresh: 0.5
best_thresh: 0.35
nms_topN_pre: 3000
nms_topN_post: 40
nms_thres: 0.4
clip_boxes: False
test_protocol: kitti
test_db: kitti
test_min_h: 0
min_det_scales: [0, 0]
cluster_anchors: 0
even_anchors: 0
expand_anchors: 0
anchors: [[-0.5, -8.5, 15.5, 23.5, 51.969, 0.531,
1.713, 1.025, -0.799],
[-8.5, -8.5, 23.5, 23.5, 52.176, 1.618,
1.6, 3.811, -0.453],
[-16.5, -8.5, 31.5, 23.5, 48.334,
1.644, 1.529, 3.966, 0.673],
[-2.528, -12.555, 17.528, 27.555,
44.781, 0.534, 1.771, 0.971, 0.093],
[-12.555, -12.555, 27.555, 27.555,
44.704, 1.599, 1.569, 3.814, -0.187],
[-22.583, -12.555, 37.583, 27.555,
43.492, 1.621, 1.536, 3.91, 0.719],
[-5.069, -17.638, 20.069, 32.638,
34.666, 0.561, 1.752, 0.967, -0.384],
[-17.638, -17.638, 32.638, 32.638,
35.35, 1.567, 1.591, 3.81, -0.511],
[-30.207, -17.638, 45.207, 32.638,
37.128, 1.602, 1.529, 3.904, 0.452],
[-8.255, -24.01, 23.255, 39.01, 28.771,
0.613, 1.76, 0.98, 0.067],
[-24.01, -24.01, 39.01, 39.01, 28.331,
1.543, 1.592, 3.66, -0.811],
[-39.764, -24.01, 54.764, 39.01,
30.541, 1.626, 1.524, 3.908, 0.312],
[-12.248, -31.996, 27.248, 46.996,
23.011, 0.606, 1.758, 0.996, 0.208],
[-31.996, -31.996, 46.996, 46.996,
22.948, 1.51, 1.599, 3.419, -1.076],
[-51.744, -31.996, 66.744, 46.996,
25.0, 1.628, 1.527, 3.917, 0.334],
[-17.253, -42.006, 32.253, 57.006,
18.479, 0.601, 1.747, 1.007, 0.347],
[-42.006, -42.006, 57.006, 57.006,
18.815, 1.487, 1.599, 3.337, -0.862],
[-66.759, -42.006, 81.759, 57.006,
20.576, 1.623, 1.532, 3.942, 0.323],
[-23.527, -54.553, 38.527, 69.553,
15.035, 0.625, 1.744, 0.917, 0.41],
[-54.553, -54.553, 69.553, 69.553,
15.346, 1.29, 1.659, 3.083, -0.275],
[-85.58, -54.553, 100.58, 69.553,
16.326, 1.613, 1.527, 3.934, 0.268],
[-31.39, -70.281, 46.39, 85.281,
12.265, 0.631, 1.747, 0.954, 0.317],
[-70.281, -70.281, 85.281, 85.281,
11.878, 1.044, 1.67, 2.415, -0.211],
[-109.171, -70.281, 124.171, 85.281,
13.58, 1.621, 1.539, 3.961, 0.189],
[-41.247, -89.994, 56.247, 104.994,
9.932, 0.61, 1.771, 0.934, 0.486],
[-89.994, -89.994, 104.994, 104.994,
8.949, 0.811, 1.766, 1.662, 0.08],
[-138.741, -89.994, 153.741, 104.994,
11.043, 1.61, 1.533, 3.899, 0.04],
[-53.602, -114.704, 68.602, 129.704,
8.389, 0.604, 1.793, 0.95, 0.806],
[-114.704, -114.704, 129.704, 129.704,
8.071, 1.01, 1.751, 2.19, -0.076],
[-175.806, -114.704, 190.806, 129.704,
9.184, 1.606, 1.526, 3.869, -0.066],
[-69.089, -145.677, 84.089, 160.677,
6.923, 0.627, 1.791, 0.96, 0.784],
[-145.677, -145.677, 160.677, 160.677,
6.784, 1.384, 1.615, 2.862, -1.035],
[-222.266, -145.677, 237.266, 160.677,
7.863, 1.617, 1.55, 3.948, -0.071],
[-88.5, -184.5, 103.5, 199.5, 5.189,
0.66, 1.755, 0.841, 0.173],
[-184.5, -184.5, 199.5, 199.5, 4.388,
0.743, 1.728, 1.381, 0.642],
[-280.5, -184.5, 295.5, 199.5, 5.583,
1.583, 1.547, 3.862, -0.072]]
bbox_means: [[-0.0, 0.002, 0.064, -0.093, 0.011,
-0.067, 0.192, 0.059, -0.021, 0.069,
-0.004]]
bbox_stds: [[0.14, 0.126, 0.247, 0.239, 0.163,
0.132, 3.621, 0.382, 0.102, 0.503,
1.855]]
anchor_scales: [32.0, 40.11, 50.276, 63.019, 78.991,
99.012, 124.106, 155.561, 194.989,
244.409, 306.354, 384.0]
anchor_ratios: [0.5, 1.0, 1.5]
hard_negatives: True
focal_loss: 0
cls_2d_lambda: 1
iou_2d_lambda: 1
bbox_2d_lambda: 0
bbox_3d_lambda: 1
bbox_3d_proj_lambda: 0.0
hill_climbing: True
visdom_port: 8100
}
Traceback (most recent call last):
File "scripts/train_rpn_3d.py", line 196, in
main(sys.argv[1:])
File "scripts/train_rpn_3d.py", line 122, in main
cls, prob, bbox_2d, bbox_3d, feat_size = rpn_net(images)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/parallel/data_parallel.py", line 159, in forward
return self.module(*inputs[0], **kwargs[0])
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/chinmay/Desktop/M3D-RPN/output/kitti_3d_multi_warmup/densenet121_3d_dilate.py", line 83, in forward
x = self.base(x)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/modules/container.py", line 117, in forward
input = module(input)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torchvision/models/densenet.py", line 111, in forward
new_features = layer(features)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torchvision/models/densenet.py", line 84, in forward
bottleneck_output = self.bn_function(prev_features)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torchvision/models/densenet.py", line 41, in bn_function
bottleneck_output = self.conv1(self.relu1(self.norm1(concated_features))) # noqa: T484
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/modules/batchnorm.py", line 131, in forward
return F.batch_norm(
File "/home/chinmay/anaconda3/lib/python3.8/site-packages/torch/nn/functional.py", line 2056, in batch_norm
return torch.batch_norm(
RuntimeError: CUDA out of memory. Tried to allocate 22.00 MiB (GPU 0; 5.79 GiB total capacity; 4.60 GiB already allocated; 3.81 MiB free; 4.72 GiB reserved in total by PyTorch)
Visualize bbox3d
Hi,
I am writing the code for visualizing both GT and predict 3d BBox on validation set images by multiplying the output 3d coordinates with the corresponding calibration matrix. Your output matches the GT, but both of those 2 projected bboxes are a bit off to the actual car. What the information I am missing? Here is the example I have.
Best,
Shukai
How to generate 3D anchors when testing?
can't convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.
m3d_rpn/lib/loss/rpn_3d.py", line 125, in forward
src_anchors = self.anchors[rois[:, 4].type(torch.cuda.LongTensor), :]
File "/home/dev-aug/.local/lib/python3.6/site-packages/torch/tensor.py", line 412, in array
return self.numpy()
TypeError: can't convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.
m3d_rpn/lib/loss/rpn_3d.py", line 202 - 213, in forward
bbox_y_tar[bind, :] = transforms[:, 1]
File "/home/dev-aug/.local/lib/python3.6/site-packages/torch/tensor.py", line 414, in array
return self.numpy().astype(dtype, copy=False)
TypeError: can't convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.
m3d_rpn/lib/loss/rpn_3d.py", line 288, in forward
src_anchors = self.anchors[rois[fg_inds, 4].type(torch.cuda.LongTensor), :]
File "/home/dev-aug/.local/lib/python3.6/site-packages/torch/tensor.py", line 412, in array
return self.numpy()
TypeError: can't convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.
please add .cpu() after each tensor before converting them to numpy array.
Question about weight decay
Hello! Could I ask a question about the hyper-parameter weight decay?
In the paper, Sec 3.5, the weight decay is 0.9.
But in the code, the weight decay is set to be 0.0005.
So which one do you actually use during training?
Thank you very much!
How to plot detection
How to plot detection from txt label file
Have you tried to train only on the car category?
Hi, I wonder that if M3D-RPN is trained only on the car category, how much will the performance increase? Have you performed this experiment? Thank you.
About the new ECCV paper
Thanks for you great work and I have been following your works for a while.
I am a bit curious about some issues about your ECCV new paper.
It seems that video prediction requires ground-truth ego-motion of the camera, at least in the training set. But I am not sure whether KITTI provides such ground truth labels for the detection dataset?
Where could I find it? Sorry for such a naive question.
The results are way better than the ones reported in the paper.
Hi Garrick, I run your code on KITTI val1 and val2 dataset. However, the results are way better than the ones reported in the paper, especially for Car. Do you have any similar results?
Val1:
Val2:
Reported in the paper:
change densenet121
thanks for your great work!But i meet a problem and i want to get your help!When i change the backbone from densenet121.features to vovnet57 or resnet50.My testing results are really bad which almost 0.And if i don't change the lr from 0.004 to smaller,it will show as below.
iter: 250, acc (bg: 0.96, fg: 0.00, iou: 0.31), loss (bbox_3d: 10.6909, cls: 0.2957, iou: 1.1781), misc (ry: 4.16, z: 7.97), dt: 0.86, eta: 11.9h
iter: 500, acc (bg: 1.00), loss (cls: 0.0065), dt: 0.80, eta: 11.1h
iter: 750, acc (bg: 1.00, fg: 0.00, iou: 0.25), loss (bbox_3d: 37.9193, cls: 0.0167, iou: 1.4007), misc (ry: 9.57, z: 24.60), dt: 0.84, eta: 11.5h
Some information losses.
i want to know if you did some ecxperiments to change the backbone.
Thank you
subprocess.CalledProcessError: Command '[...]' returned non-zero exit status 127
I tried to run this command to test the model:
python scripts/test_rpn_3d.py
and it works well until
subprocess.CalledProcessError: Command '['/home/gkd/PYTHONPROJECTServer/M3D-RPN- master/data/kitti_split1/devkit/cpp/evaluate_object', 'output/tmp_results']' ret urned non-zero exit status 127.
I traceback the code while finding the error coming from "lib/rpn_util.py"
with open(os.devnull, 'w') as devnull:
out = subprocess.check_output([script, results_path.replace('/data', '')], stderr=devnull)
how can i modify it to get the right result?
thx
about libcudart .so.9.0?
train the model with custom dataset(different size of image)
I'd like to train m3d-model using my custom dataset.I rewrite the imdb_util.py and it can generate imdb rightly.However,my image(W=3384,H=2710) in the custom dataset is larger than images in KITTI. At the beginning, I preprocess my images by cropping the top half of it,which is the part I can ignore.Now ,the image becomes W=3384,H=1855.Then,I changed the config file,using the parameters as below:
conf.test_scale = 1344 conf.crop_size = [1344, 3200]
However,it reports run out of CUDA.It seems that this size is too big to run on my gpu.
Then ,I changed the parameters smaller,like conf.test_scale= 1344 conf.crop_size =[832, 2048],
but I got acc (bg: 1.00, fg: 0.00, iou: nan), loss (bbox_3d: inf, cls: 2000.0002, iou: nan), misc (ry: inf, z: inf), dt: 1.70, eta: 20.7h .Is it because of the unfittable initialization of anchor size ,affected by crop_size?
So what should I do to train my dataset rightly without the development of gpu?
I'd really appreciate it if you could do me a favor!
How can I train the model on my own dataset?
Hi, @garrickbrazil, M3D is a really great work in the monocular vision. But how can I train the model on my own dataset, and the image size is different from KITTI image size.
How can i use the trained weight to test one single image and plot the result 3D boxes in the picture
and by the way, what's differences between model_pkl and optim_pkl
thanks~
The nms problem
I test the code,it seems that nms does not work.The effect is shown in the figure below：
and can not detect object in far distance.Is anybody know how to fix it? Thank you!
M3D-RPN-master/lib/imdb_util.py", line 336, in read_kitti_cal return p2 UnboundLocalError: local variable 'p2' referenced before assignment
when runnning "python scripts/train_rpn_3d.py --config=kitti_3d_multi_warmup", it met the error "M3D-RPN-master/lib/imdb_util.py", line 336, in read_kitti_cal return p2 UnboundLocalError: local variable 'p2' referenced before assignment"
Some question in read_kitti_label.py
Thanks to the author for releasing the code and I have some questions about code details which are really confused me. Why do u recode cy3d -= (h3d / 2), rotY, x, y, x2, y2 in read_kitti_label and what are they means after corresponding operations?
depth-awre convolution does not work well
When I trained the model with and without depth-aware convolution on val1, I found that there is no significant difference between them. The results are what we trained without depth-aware:
test_iter 50000 2d car --> easy: 0.9041, mod: 0.8394, hard: 0.6780
test_iter 50000 gr car --> easy: 0.2736, mod: 0.2127, hard: 0.1770
test_iter 50000 3d car --> easy: 0.2100, mod: 0.1702, hard: 0.1522
test_iter 50000 2d pedestrian --> easy: 0.6733, mod: 0.5882, hard: 0.5056
test_iter 50000 gr pedestrian --> easy: 0.0689, mod: 0.0677, hard: 0.0646
test_iter 50000 3d pedestrian --> easy: 0.0397, mod: 0.0385, hard: 0.0328
test_iter 50000 2d cyclist --> easy: 0.6383, mod: 0.4694, hard: 0.4062
test_iter 50000 gr cyclist --> easy: 0.0545, mod: 0.0563, hard: 0.0565
test_iter 50000 3d cyclist --> easy: 0.0432, mod: 0.0555, hard: 0.0557
And the results we trained with depth-aware is:
test_iter 50000 2d car --> easy: 0.9210, mod: 0.8443, hard: 0.6824
test_iter 50000 gr car --> easy: 0.2716, mod: 0.2195, hard: 0.1789
test_iter 50000 3d car --> easy: 0.2136, mod: 0.1745, hard: 0.1554
test_iter 50000 2d pedestrian --> easy: 0.6637, mod: 0.5836, hard: 0.5011
test_iter 50000 gr pedestrian --> easy: 0.0583, mod: 0.0561, hard: 0.0492
test_iter 50000 3d pedestrian --> easy: 0.0565, mod: 0.0529, hard: 0.0439
test_iter 50000 2d cyclist --> easy: 0.7268, mod: 0.4851, hard: 0.4810
test_iter 50000 gr cyclist --> easy: 0.0712, mod: 0.0584, hard: 0.0561
test_iter 50000 3d cyclist --> easy: 0.0594, mod: 0.0549, hard: 0.0551
As the Tab. 5 in the paper, there should be a significant improvement in 3D and BEV performance with depth-aware (about 6% and 8%) compared to that without depth-aware, but we do not find that (only about 0.4% and 0.7%). And we also notice that the performance of pedestrain and cyclist is also not as good, only 5% for pedestrain and cyclist, but in Tab. 3, it is about 10%.
Anchors and how to draw them on an image
Dear @garrickbrazil ,
I have a question regarding the anchors. In your paper, there is a part, where you have displayed shown the anchors, 12 of them to be exact. I know that the anchors, total of 36, are calculated via "generate_anchors(conf, imdb, cache_folder)"
Line 22 in 05ef276
each anchor has 9 dimensions. Could you please describe what they are? and how can I display the anchors i.e. draw them on an image? I really appreciate your feedback regarding this matter.
Best regards and thanks in advance,
Sam
projection_ray_trace
Hi,
It is a nice work!!! Thanks for releasing the source code.
When I study your code, I find there is a function named projection_ray_trace in the file imdb_util.py, but I do not find the definition, can you check it?
Thanks a lot.
TypeError: can't convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.
python scripts/train_rpn_3d.py --config=kitti_3d_multi_warmup:
Traceback (most recent call last):
File "scripts/train_rpn_3d.py", line 196, in
main(sys.argv[1:])
File "scripts/train_rpn_3d.py", line 125, in main
det_loss, det_stats = criterion_det(cls, prob, bbox_2d, bbox_3d, imobjs, feat_size)
File "/home/user/anaconda3/envs/M3DRPN/lib/python3.6/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/home/user/yty/M3D-RPN-master/lib/loss/rpn_3d.py", line 126, in forward
src_anchors = self.anchors[rois[:, 4].type(torch.cuda.LongTensor), :]
File "/home/user/anaconda3/envs/M3DRPN/lib/python3.6/site-packages/torch/tensor.py", line 486, in array
return self.numpy()
TypeError: can't convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.
about the data
i have noted that a directory 'kitti_raw['pre'] = os.path.join(base_data, 'kitti', 'training', 'prev_2')' was required when prepare the kitti kitti_split1 dataset in M3D-RPN-master/data/kitti_split1/setup_split.py. why this is need in the project? looking forward to your reply
Academic fraud, code totally wrong
Are u sure the result can be reproduce?
Are u sure this code can be run? Not mentation there were fuction missing in other issue, but this code is mess and totally wrong:
src_anchors = self.anchors[rois[:, 4].cpu().numpy(), :]
rois is float, how do u able to index an array with float????
Training without warmup
Thanks for your great work!
I have a question about training.
Do I have to train with warmup config first?
Or I could directly train with main config without warmup?
And what's the function of warmup training?
Thanks:)
prepare datasets
sorry,i click the "kitti" link.But there are many links:
Download left color images of object data set (12 GB)
Download right color images, if you want to use stereo information (12 GB)
Download the 3 temporally preceding frames (left color) (36 GB)
Download the 3 temporally preceding frames (right color) (36 GB)
Download Velodyne point clouds, if you want to use laser information (29 GB)
Download camera calibration matrices of object data set (16 MB)
Download training labels of object data set (5 MB)
Download object development kit (1 MB) (including 3D object detection and bird's eye view evaluation code)
Download pre-trained LSVM baseline models (5 MB) used in Joint 3D Estimation of Objects and Scene Layout (NIPS 2011). These models are referred to as LSVM-MDPM-sv (supervised version) and LSVM-MDPM-us (unsupervised version) in the tables below.
Download reference detections (L-SVM) for training and test set (800 MB)
Qianli Liao (NYU) has put together code to convert from KITTI to PASCAL VOC file format (documentation included, requires Emacs).
Karl Rosaen (U.Mich) has released code to convert between KITTI, KITTI tracking, Pascal VOC, Udacity, CrowdAI and AUTTI formats.
datasets in this link page all needed to be downloaded?
About KITTI data folder
Hi Author, i found that four folders are used inside kitti folder like (calib,image_2,label_2,prev_2). i got the calibration matrix, images and labels. i am confused what is prev_2 in this dataset. kindly can you explain it.
thanks in advance.
No module named 'lib.imdb_util'
No module named 'lib.imdb_util'
When I do python scripts/test_rpn_3d.py
runtime error
I got an error with the following message, I don not known how to solve it, could you help me?
Traceback (most recent call last):
File "scripts/train_rpn_3d.py", line 197, in
main(sys.argv[1:])
File "scripts/train_rpn_3d.py", line 123, in main
cls, prob, bbox_2d, bbox_3d, feat_size = rpn_net(images)
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 547, in call
result = self.forward(*input, **kwargs)
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 153, in forward
return self.gather(outputs, self.output_device)
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 165, in gather
return gather(outputs, output_device, dim=self.dim)
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/scatter_gather.py", line 68, in gather
res = gather_map(outputs)
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
return type(out)(map(gather_map, zip(*outputs)))
TypeError: zip argument #1 must support iteration
What does the time mean in the paper?
When I read the paper, I have a confusion that what is the time in below chart represents?
Does the 161 ms here including the preprocess of image and post process of pair constraints by g2o?
Or it only represents the time of network flow?
Hoping for your reply.Thanks
How is the performace when using resnet( 18 or 50)
I just want to know which backbone is good enough to train for a mono 3d prediction task.
How can i learn only "Pedestrian"?
i wanted to learn only Pedestrian ,so i changed config file.
conf.lbls = [ 'Pedestrian','Car', 'Cyclist'] ---->conf.lbls = [ 'Pedestrian']
But it make an errorraise ValueError('Non-used anchor #{} found'.format(aind))
how can i lean only "Pedestrian"?
how to detect and plot anchors in new image
hello I am newer in 3D object detection, soory if my question is native, i retrained the model to detect cars only via colab ,and now I have model_50000_pkl , and conf.pkl ,my question is how to use it to detect and visualize the result in a new image, I really need an answer,thank you very much
raise ValueError('Non-used anchor #{} found'.format(aind))
Hello, I converted lyft open dataset to Kitti format and tried to train on the converted dataset, but I got error "raise ValueError('Non-used anchor #{} found'.format(aind))" on rpn_util.py, after some debugging, I found the problem is that some anchors were never chosen as max prob target anchor, so I want to know how to adjust hyper parameters to fix the problem, I think it's quite tricky to choose some parameters that can make each anchor has probability to be the max prob target anchor, and how these anchors will affect the performance. Hope to get some ideas from you or guys here
Data split
Hi @garrickbrazil,
Thanks for making this work public. I wanted to understand more about the data split, why are they split in particular way (i.e., if they are following any guidelines etc) and are there 2 versions of data split?
Would really appreciate if you could provide the reason for choosing smaller training dataset than validation set (3682 training & 3799 val)
Thanks
reimplement in google colab
Dear authors,
I notice that all the instructions for reimplementation in readme are based on local environment.
However, if I am wanna reimplement in google colab, is there any guideline or anything that I need to notice?
Thanks!
About split dataset:"test"
The paper told that there are three datasets in the evaluation: kitti_split1(val1),kitti_split2(val2) and test .
My question is that test dataset seems to use the kitti/training to train and use kitti_split1/val1 to test.
Won't it be some overlapping between the test part and the train part,which runs counter to the paper.
And paper told me that it used official test dataset,but when I print out the test_model,it shows
dataset_test: kitti_split1 datasets_train: [{'anno_fmt': 'kitti_det', 'im_ext': '.png', 'name': 'kitti', 'scale': 1}]
so what does "test" actually mean in the paper?
About loss setting
Hi @garrickbrazil
Nice work~ I have two questions.
-
Loss. In the scripts/config/kitti_3d_multi_main.py, the loss weight of bbox_3d_proj_lambda and bbox_2d_lambda is set to 0. Why do you discard the two losses？
-
pretrained model. Could you offer your pretrained two models (warmup and main). We are trying to convert it into another DL framework.
Hope you can help us, thank you!
visdom not show any picture or plot when training
hi,Mr Garrick Brazil .thanks for your work.when i try reproduce the results,some problems came in traing stage.
ubuntu 16.04
cuda 8.0 cudnn 6.0
pytorch:1.0.1.post2
gpu:1080ti
firstly: python -m visdom.server -port 8100 -readonly
when i run command:python scripts/train_rpn_3d.py --config=kitti_3d_multi_warmup
machine training works
but visdom does not show any after 1hour,
Mr Garrick Brazil,may the problem caused by datasets?
train the model using a model_20000_pkl error
hello, i sorry to interrupt you, i meet a error when i try to run the script of train_rpn_3d.py with a pretrained weights model_20000_pk.the weights pkl file is from a warmup training. the error is :File "/home/PycharmProjects/models_git/M3D-RPN-master/scripts/../../M3D-RPN-master/output/kitti_3d_multi_main/densenet121_3d_dilate_depth_aware.py", line 302, in build
src_weights[dst_weight_key] = src_weights[src_weight_key].repeat(conf.bins, 1, 1, 1)
KeyError: 'module.prop_feats.0.weight'. seems that i can not use the pretrained model,
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.