main.py error: argument --nThread: expected one argument

[kunfei1]

Hellow, I used GX1080*1 8G and ran demo.sh. Got this problem: usage: main.py [-h] [--nThread NTHREAD] [--cpu] [--nGPU NGPU] [--datadir DATADIR] [--data_train DATA_TRAIN] [--data_test DATA_TEST] [--reset] [--epochs EPOCHS] [--test_every TEST_EVERY] [--batchid BATCHID] [--batchimage BATCHIMAGE] [--batchtest BATCHTEST] [--test_only] [--model MODEL] [--loss LOSS] [--act ACT] [--pool POOL] [--feats FEATS] [--height HEIGHT] [--width WIDTH] [--num_classes NUM_CLASSES] [--lr LR] [--optimizer {SGD,ADAM,ADAMAX,RMSprop}] [--momentum MOMENTUM] [--dampening DAMPENING] [--nesterov] [--beta1 BETA1] [--beta2 BETA2] [--amsgrad] [--epsilon EPSILON] [--gamma GAMMA] [--weight_decay WEIGHT_DECAY] [--decay_type DECAY_TYPE] [--lr_decay LR_DECAY] [--margin MARGIN] [--re_rank] [--random_erasing] [--probability PROBABILITY] [--savedir SAVEDIR] [--outdir OUTDIR] [--resume RESUME] [--save SAVE] [--load LOAD] [--save_models] [--pre_train PRE_TRAIN] main.py: error: argument --nThread: expected one argument

What can I do to solve it?

[seathiefwang]

Only one gpu? modify demo.sh CUDA_VISIBLE_DEVICES=0,1 python3 main.py --datadir Market1501_dir --reset --re_rank --random_erasing --margin 1.2 --optimizer ADAM --epochs 160 --decay_type step_120_140 --test_every 20 --nGPU 2 --save MGN_save

to

python3 main.py --datadir Market1501_dir --reset --re_rank --random_erasing --margin 1.2 --optimizer ADAM --epochs 160 --decay_type step_120_140 --test_every 20 --nGPU 1 --nThread 2 --save MGN_save

[kunfei1]

thank you for your help. when worked , it show: [INFO] [41/160] 94/94 [CrossEntropy: 4.5233][Triplet: 0.3194][Total: 4.8427] what the 94 means? the batchid id set 8 and the batchimage set 8, used market-1501

[seathiefwang]

total batch.

[kunfei1]

if that ,only 94*16=1504 has used, but the market has 12937 imges, most of them been abandoned？

[seathiefwang]

94 8 8=6016, each id selects 8(batchimage) pictures for training, and others will be discarded, every epoch will be random select.

[chenshuo0012]

hello,when I run you code at my 1080 use python3 main.py --datadir Market1501_dir --reset --re_rank --random_erasing --margin 1.2 --optimizer ADAM --epochs 160 --decay_type step_120_140 --test_every 20 --nGPU 1 --nThread 2 --save MGN_save the error occured: Traceback (most recent call last): File "main.py", line 13, in model = model.Model(args, ckpt) File "/home/cs/MGN-pytorch-master/model/init.py", line 28, in init print(self.model, file=ckp.log_file) NameError: name 'ckp' is not defined What can I do to solve it?

[chenshuo0012]

I slove the problem, and the problem of out of memory. thank you very much for your share.

[kunfei1]

could you convenient to tell the modified parameter in the last model? I only trained it at 94.49%, and couled not improve it

[seathiefwang]

@kunfei1 python3 main.py --datadir Market-1501-v15.09.15 --batchid 16 --batchtest 32 --test_every 100 --epochs 300 --decay_type step_250_290 --loss 1CrossEntropy+2Triplet --margin 1.2 --save adam_3 --nGPU 2 --lr 2e-4 --optimizer ADAM --random_erasing --reset --re_rank --amsgrad

[chj1933]

@seathiefwang , What is the function of the (--decay_type step_250_290) ?

[seathiefwang]

Reduce learning rate in the 250th epoch and 290th epoch.

[Jinjicheng]

@seathiefwang ,thank you for sharing your code. i run your code with demo.sh, mAP is 0.8056,it is lower than your,why and how to improve it?

[chj1933]

@seathiefwang ，Hellow, I Have a question, If I want to use each id all pictures for training instead of selecting 8(batchimage) pictures for training,, How should I set parameters？

[seathiefwang]

@chj1933 Each epoch will be randomly selected and all images will be used.

[seathiefwang]

@chj1933 I updated demo.sh file.

[chapmancpp]

@seathiefwang ,老师我用自己的pc跑的，除了batchid 4以外，其他的都是您的参数，[CrossEntropy: 0.5297][Triplet: 0.0000]。损失到这边就下不去了，请问是我batchid大小的问题吗。精度到82%就上不去了，测试了一下您训练好的模型，我这边精度只有87.5%。是我有什么步骤没做吗？

[chj1933]

@seathiefwang ，你好，老师！你的MGN代买除了能测试single-shot，能不能测试multi-shot?