METEOR in eval.py about paperrobot HOT 4 CLOSED

I think you should run it under ubuntu. The meteor package might not function well under windows. Also I recommend using python3.6. Actually, you can comment out the eval.py Line 15 and line 72, since we don’t use meteor in train. On Tue, Jul 9, 2019 at 7:58 AM demonlj <[email protected]<mailto:[email protected]>> wrote: got similar error system: win8.1 + cuda10.1 + python 3.7 + pytorch1.1 C:\greenSoft\PaperRobot-master\New paper writing>python train.py --data_path dat a/pubmed_abstract --model_dp abstract_model/ --batch_size=10 `Model location: abstract_model/memory,lower=False,freq=5,word_dim=128,h=8,hop=3, input_dropout_p=0.2,layer_dropout=0.2,gpu=True,batch_size=10,max_len=150,lr_meth od=adam,lr_rate=0.001 Model parameters: lower=False freq=5 word_dim=128 h=8 hop=3 pre_emb= input_dropout_p=0.2 layer_dropout=0.2 gpu=True batch_size=10 max_len=150 lr_method=adam lr_rate=0.001 data_path=data/pubmed_abstract Found 25133 unique words (16508545 in total) 22811 / 2095 pairs in train / dev. train batches 2282 Epoch 0/99 2281 batches processed. current batch loss: 6.061568 Finished epoch 0: Train Avg NLLLoss: 6.4136 Traceback (most recent call last): File "train.py", line 261, in eval_f = Evaluate() File "C:\greenSoft\PaperRobot-master\New paper writing\eval.py", line 15, in _ init_ (Meteor(), "METEOR"), File "C:\greenSoft\PaperRobot-master\New paper writing\pycocoevalcap\meteor\me teor.py", line 20, in init stdout=subprocess.PIPE, stderr=subprocess.PIPE, universal_newlines=True, buf size=1) File "C:\Users\MyPC\AppData\Local\Programs\Python\Python37\lib\subprocess.py", line 775, in init restore_signals, start_new_session) File "C:\Users\MyPC\AppData\Local\Programs\Python\Python37\lib\subprocess.py", line 1178, in _execute_child startupinfo) FileNotFoundError: [WinError 2]。 Exception ignored in: <function Meteor.del at 0x0000000FECFBDD90> Traceback (most recent call last): File "C:\greenSoft\PaperRobot-master\New paper writing\pycocoevalcap\meteor\me teor.py", line 59, in del self.lock.acquire() AttributeError: 'Meteor' object has no attribute 'lock' ` Thought the training already finished, the METEOR error does not matter? — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub<#4?email_source=notifications&email_token=ADNENRPMAOSNDUXS73CZSETP6PIA7A5CNFSM4HYNR7QKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZOV3IQ#issuecomment-509435298>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ADNENRP3NWRLRP3I66STXK3P6PIA7ANCNFSM4HYNR7QA>.