FriendChi

当我启用了SegHead结构后，就发生了报错。

2

Traceback (most recent call last): File "train.py", line 142, in train(net, train_loader, loss_dict, optimizer, scheduler,logger, epoch, metric_dict, cfg.dataset) File "train.py", line 30, in train loss = calc_loss(loss_dict, results, logger, global_step,...

Why does it perform so poorly in the test7_cross.txt, is this normal？

2

38248.1s | 4060 | ------------Configuration--------- 38248.1s | 4061 | anno_dir: /kaggle/input/culane/CULane/ 38248.1s | 4062 | detect_dir: /kaggle/working/Ultra-Fast-Lane-Detection-v2/log/20240610_063247_lr_6e-03_b_32/culane_eval_tmp/ 38248.1s | 4063 | im_dir: /kaggle/input/culane/CULane/ 38248.1s | 4064 | list_im_file: /kaggle/working/temp/list/list/test_split/test7_cross.txt 38248.1s...

训练时加载数据killed，是不是对内存要求相当高？

1

root@81c1814922ec:/app# python main.py data/730-25 --workspace model/trial_730-25 -O --iters 100000 --finetune_lips --patch_size 64 --asr_model ave --preload 2 Namespace(H=450, O=True, W=450, amb_aud_loss=1, amb_dim=2, amb_eye_loss=1, asr=False, asr_model='ave', asr_play=False, asr_save_feats=False, asr_wav='', att=2, au45=False, aud='',...

使用python scripts/convert_curvelanes.py出现了下面的警告 Ultra-Fast-Lane-Detection-v2/scripts/convert_curvelanes.py:31: RankWarning: Polyfit may be poorly conditioned curve1 = np.polyfit(line_y[:2], line_x[:2], deg = 1) Ultra-Fast-Lane-Detection-v2/scripts/convert_curvelanes.py:71: RankWarning: Polyfit may be poorly conditioned curve = np.polyfit(arr[1::2], arr[::2], min(len(arr[::2]) - 1,...