rsennrich
commented
3 years ago Hello Amit,
it looks like you don't have the permission to execute "$script_dir/validate.sh". Can you check the file permissions? If you let me know how you downloaded the files, I can also try to fix this upstream. If I clone https://github.com/EdinburghNLP/wmt17-scripts , permissions seem to be correct, though.
best wishes, Rico
On 02/04/2021 07:43, AMIT KUMAR wrote:
Here is the training file
!/bin/sh
Distributed under MIT license
script_dir=|dirname $0| main_dir=$script_dir/../ data_dir=$main_dir/data working_dir=$main_dir/model
device=0,1
independent variables (toolkit locations)
. $main_dir/../vars
language-dependent variables (source and target language)
. $main_dir/vars
CUDA_VISIBLE_DEVICES=$device python $nematus_home/nematus/train.py --model $working_dir/model --datasets $data_dir/korean-english-park.train.bpe.$src $data_dir/korean-english-park.train.bpe.$trg --valid_datasets $data_dir/korean-english-park.dev.bpe.$src $data_dir/korean-english-park.dev.bpe.$trg --dictionaries $data_dir/korean-english-park.train.bpe.ko.json $data_dir/korean-english-park.train.bpe.en.json --valid_script $script_dir/validate.sh --reload latest_checkpoint --dim_word 512 --dim 1024 --lrate 0.0005 --optimizer adam --maxlen 200 --batch_size 80 --valid_batch_size 40 --validFreq 10000 --dispFreq 1000 --saveFreq 30000 --sampleFreq 10000 --tie_decoder_embeddings --layer_normalisation --dec_base_recurrence_transition_depth 2 --enc_recurrence_transition_depth 2 --rnn_dropout_hidden 0.5 --rnn_dropout_embedding 0.5 --rnn_dropout_source 0.3 --rnn_dropout_target 0.3 --label_smoothing 0.2 --token_batch_size 1000 --valid_token_batch_size 1000 --patience 10
validate.sh file mentioned below:
!/bin/sh
Distributed under MIT license
this script evaluates translations of the newstest2013 test set
using detokenized BLEU (equivalent to evaluation with mteval-v13a.pl).
script_dir=|dirname $0| main_dir=$script_dir/../ data_dir=$main_dir/data working_dir=$main_dir/model
device=0,1
language-independent variables (toolkit locations)
. $main_dir/../vars
language-dependent variables (source and target language)
. $main_dir/vars
dev_prefix=korean-english-park.dev dev=$dev_prefix.bpe.$src ref=$dev_prefix.$trg prefix=$working_dir/model
decode
for new Tensorflow backend, use a command like this:
CUDA_VISIBLE_DEVICES=$device python $nematus_home/nematus/translate.py -m $prefix -i $data_dir/$dev -o $working_dir/$dev.output.dev -k 5 -n
bash $script_dir/postprocess.sh < $working_dir/$dev.output.dev > $working_dir/$dev.output.postprocessed.dev
get BLEUBEST=|cat ${prefix}_best_bleu || echo 0| $nematus_home/data/multi-bleu-detok.perl /scratch/amitkumar.rs.cse17.iitbhu/ko-en/parallel/$ref < $working_dir/$dev.output.postprocessed.dev >> ${prefix}_bleu_scores BLEU=|$nematus_home/data/multi-bleu-detok.perl /scratch/amitkumar.rs.cse17.iitbhu/ko-en/parallel/$ref < $working_dir/$dev.output.postprocessed.dev | cut -f 3 -d ' ' | cut -f 1 -d ','| BETTER=|echo "$BLEU > $BEST" | bc|
echo "BLEU = $BLEU"
echo "BLEU = $BLEU"
if [ "$BETTER" = "1" ]; then echo "new best; saving" echo $BLEU > ${prefix}_best_bleu cp ${prefix}.json ${prefix}.best_bleu.json fi
Following errors in training:
2021-04-02 07:11:20.787601: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 INFO: Namespace(adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08, batch_size=80, beam_freq=10000, beam_size=12, clip_c=1.0, datasets=['scripts/..//data/korean-english-park.train.bpe.ko', 'scripts/..//data/korean-english-park.train.bpe.en'], decay_c=0.0, dictionaries=['scripts/..//data/korean-english-park.train.bpe.ko.json', 'scripts/..//data/korean-english-park.train.bpe.en.json'], dim_per_factor=[512], disp_freq=1000, embedding_size=512, exponential_smoothing=0.0, factors=1, finish_after=10000000, gradient_aggregation_steps=1, keep_train_set_in_memory=False, label_smoothing=0.2, layer_normalization_type='layernorm', learning_rate=0.0005, learning_schedule='constant', loss_function='cross-entropy', map_decay_c=0.0, max_epochs=5000, max_len_a=1.5, max_len_b=5, max_sentences_of_sampling=0, max_sentences_per_device=0, max_tokens_per_device=0, maxibatch_size=20, maxlen=200, model_type='rnn', model_version=0.2, mrt_alpha=0.005, mrt_loss='SENTENCEBLEU n=4', mrt_ml_mix=0, mrt_reference=False, n_best=False, normalization_alpha=0.0, optimizer='adam', output_hidden_activation='tanh', patience=10, plateau_steps=0, preprocess_script=None, print_per_token_pro=False, prior_model=None, reload='latest_checkpoint', reload_training_progress=True, rnn_dec_base_transition_depth=2, rnn_dec_deep_context=False, rnn_dec_depth=1, rnn_dec_high_transition_depth=1, rnn_dropout_embedding=0.5, rnn_dropout_hidden=0.5, rnn_dropout_source=0.3, rnn_dropout_target=0.3, rnn_enc_depth=1, rnn_enc_transition_depth=2, rnn_layer_normalization=True, rnn_lexical_model=False, rnn_use_dropout=True, sample_freq=10000, sample_way='beam_search', samplesN=100, sampling_temperature=1.0, save_freq=30000, saveto='scripts/..//model/model', shuffle_each_epoch=True, softmax_mixture_size=1, sort_by_length=True, source_dataset='scripts/..//data/korean-english-park.train.bpe.ko', source_dicts=['scripts/..//data/korean-english-park.train.bpe.ko.json'], source_vocab_sizes=[9910], state_size=1024, summary_dir=None, summary_freq=0, target_dataset='scripts/..//data/korean-english-park.train.bpe.en', target_dict='scripts/..//data/korean-english-park.train.bpe.en.json', target_embedding_size=512, target_vocab_size=5953, theano_compat=False, tie_decoder_embeddings=True, tie_encoder_decoder_embeddings=False, token_batch_size=1000, transformer_dec_depth=6, transformer_drophead=0.0, transformer_dropout_attn=0.1, transformer_dropout_embeddings=0.1, transformer_dropout_relu=0.1, transformer_dropout_residual=0.1, transformer_enc_depth=6, transformer_ffn_hidden_size=2048, transformer_num_heads=8, translation_maxlen=200, translation_strategy='beam_search', valid_batch_size=40, valid_bleu_source_dataset='scripts/..//data/korean-english-park.dev.bpe.ko', valid_datasets=['scripts/..//data/korean-english-park.dev.bpe.ko', 'scripts/..//data/korean-english-park.dev.bpe.en'], valid_freq=10000, valid_script='scripts/validate1.sh', valid_source_dataset='scripts/..//data/korean-english-park.dev.bpe.ko', valid_target_dataset='scripts/..//data/korean-english-park.dev.bpe.en', valid_token_batch_size=1000, warmup_steps=8000) 2021-04-02 07:11:41.096118: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN)to use the following CPU instructions in performance-critical operations: AVX2 AVX512F FMA To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags. 2021-04-02 07:11:41.108886: I tensorflow/core/platform/profile_utils/cpu_utils.cc:104] CPU Frequency: 2400000000 Hz 2021-04-02 07:11:41.110970: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x5612befedba0 initialized for platform Host (this does not guarantee that XLA will be used). Devices: 2021-04-02 07:11:41.111051: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version 2021-04-02 07:11:41.116171: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcuda.so.1 2021-04-02 07:11:42.267591: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x5612bf0af6e0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices: 2021-04-02 07:11:42.267741: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla V100-PCIE-16GB, Compute Capability 7.0 2021-04-02 07:11:42.270571: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1716] Found device 0 with properties: pciBusID: 0000:5e:00.0 name: Tesla V100-PCIE-16GB computeCapability: 7.0 coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 15.75GiB deviceMemoryBandwidth: 836.37GiB/s 2021-04-02 07:11:42.270674: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 2021-04-02 07:11:42.276043: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10 2021-04-02 07:11:42.279407: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcufft.so.10 2021-04-02 07:11:42.280411: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcurand.so.10 2021-04-02 07:11:42.283420: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusolver.so.10 2021-04-02 07:11:42.285053: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusparse.so.10 2021-04-02 07:11:47.565824: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudnn.so.7 2021-04-02 07:11:47.571036: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1858] Adding visible gpu devices: 0 2021-04-02 07:11:47.571172: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 2021-04-02 07:11:49.279069: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1257] Device interconnect StreamExecutor with strength 1 edge matrix: 2021-04-02 07:11:49.279154: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1263] 0 2021-04-02 07:11:49.279172: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1276] 0: N 2021-04-02 07:11:49.281268: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1402] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 14729 MB memory) -> physical GPU (device: 0, name: Tesla V100-PCIE-16GB, pci bus id: 0000:5e:00.0, compute capability: 7.0) 2021-04-02 07:11:49.286117: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1716] Found device 0 with properties: pciBusID: 0000:5e:00.0 name: Tesla V100-PCIE-16GB computeCapability: 7.0 coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 15.75GiB deviceMemoryBandwidth: 836.37GiB/s 2021-04-02 07:11:49.286209: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 2021-04-02 07:11:49.286244: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10 2021-04-02 07:11:49.286269: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcufft.so.10 2021-04-02 07:11:49.286293: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcurand.so.10 2021-04-02 07:11:49.286316: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusolver.so.10 2021-04-02 07:11:49.286340: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusparse.so.10 2021-04-02 07:11:49.286364: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudnn.so.7 2021-04-02 07:11:49.288785: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1858] Adding visible gpu devices: 0 2021-04-02 07:11:49.288839: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1257] Device interconnect StreamExecutor with strength 1 edge matrix: 2021-04-02 07:11:49.288879: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1263] 0 2021-04-02 07:11:49.288899: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1276] 0: N 2021-04-02 07:11:49.291537: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1402] Created TensorFlow device (/device:GPU:0 with 14729 MB memory) -> physical GPU (device: 0, name: Tesla V100-PCIE-16GB, pci bus id: 0000:5e:00.0, compute capability: 7.0) INFO: Building model... WARNING: From /home/amitkumar.rs.cse17.iitbhu/nematus/nematus/rnn_model.py:59: dropout (from tensorflow.python.keras.legacy_tf_layers.core) is deprecated and will be removed in a future version. Instructions for updating: Use keras.layers.dropout instead. WARNING: From /home/amitkumar.rs.cse17.iitbhu/envNem1/lib/python3.5/site-packages/tensorflow-2.3.1-py3.5-linux-x86_64.egg/tensorflow/python/keras/legacy_tf_layers/core.py:271: Layer.apply (from tensorflow.python.keras.engine.base_layer_v1) is deprecated and will be removed in a future version. Instructions for updating: Please use |layer.call| method instead. 2021-04-02 07:11:51.977754: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1716] Found device 0 with properties: pciBusID: 0000:5e:00.0 name: Tesla V100-PCIE-16GB computeCapability: 7.0 coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 15.75GiB deviceMemoryBandwidth: 836.37GiB/s 2021-04-02 07:11:51.977826: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 2021-04-02 07:11:51.977870: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10 2021-04-02 07:11:51.977889: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcufft.so.10 2021-04-02 07:11:51.977907: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcurand.so.10 2021-04-02 07:11:51.977925: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusolver.so.10 2021-04-02 07:11:51.977942: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusparse.so.10 2021-04-02 07:11:51.977965: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudnn.so.7 2021-04-02 07:11:51.979691: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1858] Adding visible gpu devices: 0 INFO: Initializing model parameters from scratch... INFO: Done INFO: Reading data... INFO: Done INFO: Initial uidx=0 INFO: Starting epoch 0 2021-04-02 07:12:23.860174: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10 INFO: [2021-04-02 07:33:59] Epoch: 0 Update: 1000 Loss/word: 6.966987346043379 Words/sec: 696.856504668986 Sents/sec: 21.713141805022023 INFO: [2021-04-02 07:55:22] Epoch: 0 Update: 2000 Loss/word: 6.580592949380649 Words/sec: 703.13946539418 Sents/sec: 22.02518408136046 INFO: [2021-04-02 08:16:18] Epoch: 0 Update: 3000 Loss/word: 6.428336913131417 Words/sec: 722.7127947165809 Sents/sec: 22.798869988266855 INFO: Starting epoch 1 INFO: [2021-04-02 08:37:23] Epoch: 1 Update: 4000 Loss/word: 6.330259053899345 Words/sec: 716.8921478972981 Sents/sec: 22.429987003584493 INFO: [2021-04-02 08:58:25] Epoch: 1 Update: 5000 Loss/word: 6.2662326674608595 Words/sec: 717.9178678249805 Sents/sec: 22.6803643125988 INFO: [2021-04-02 09:19:13] Epoch: 1 Update: 6000 Loss/word: 6.226474244214676 Words/sec: 726.10896397206 Sents/sec: 22.971330964205684 INFO: Starting epoch 2 INFO: [2021-04-02 09:40:21] Epoch: 2 Update: 7000 Loss/word: 6.17927944352242 Words/sec: 714.946995990232 Sents/sec: 22.401948114913043 INFO: [2021-04-02 10:01:17] Epoch: 2 Update: 8000 Loss/word: 6.147852558704621 Words/sec: 719.9293130745198 Sents/sec: 22.550224623889665 INFO: [2021-04-02 10:22:19] Epoch: 2 Update: 9000 Loss/word: 6.124515647550154 Words/sec: 718.4753338888931 Sents/sec: 22.38133746941181 INFO: Starting epoch 3 INFO: [2021-04-02 10:43:16] Epoch: 3 Update: 10000 Loss/word: 6.093494990647863 Words/sec: 722.7513458365172 Sents/sec: 22.890552982759697 INFO: SOURCE: T@@ M@@ Z@@ 닷@@ 컴@@ 은 듀@@ 브@@ 로가 네@@ 버@@ 더@@ 주 라@@ 스@@ 베이@@ 거@@ 스에 있는 자택에서 사망했다고 보도했다 . INFO: TARGET: Du@@ B@@ row died at his home in Las Vegas , Nevada , according to T@@ M@@ Z@@ .@@ com@@ . INFO: SAMPLE: American G@@ reported C@@ D wasn 't Roger C@@ bad C@@ y@@ d@@ y@@ d , Paris and h@@ ol@@ ly behind your rushed safely today . INFO: SOURCE: 추@@ 후 에 요@@ 금이 적용@@ 될 수도 있지만 현@@ 재@@ 는 무@@ 료@@ 다 . INFO: TARGET: initially , there will be no f@@ ee to apply for travel author@@ iz@@ ation , although one may be imposed later . INFO: SAMPLE: after car@@ c@@ ers often be used pa@@ rent , but a gi@@ gh@@ ty ency has been thought to vote oe @-@ on test ! INFO: SOURCE: 일부 펀@@ 드@@ 매@@ 니@@ 저는 " 기업 이@@ 윤@@ 이 증가@@ 해 주식@@ 시장이 활@@ 성@@ 화@@ 됐다 " 고 전했다 . INFO: TARGET: fund man@@ ag@@ ers said strong corporate profit growth was supporting stocks . INFO: SAMPLE: more than $ 3.@@ 5 bailout attacked C@@ m@@ succeed for more barrels a rise , " 29@@ ,000 yen re@@ str@@ ong@@ s to investors . INFO: SOURCE: 이 바이러@@ 스는 서@@ 쪽으로 빠르게 퍼@@ 지고 있으며 , 관리들은 이것이 곧 서부 해안@@ 에@@ 까지 이를 것으로 우려하고 있다 . INFO: TARGET: the virus has been spreading quickly w@@ est@@ ward , with officials f@@ ear@@ ing it will soon reach the West Coast . INFO: SAMPLE: camps are cleared to sea winds and often @-@ anger gl@@ om@@ ing to Iraq as " coast of Hospital always . documentary loc@@ als politicians northwest the route . term of increasing exposure because INFO: SOURCE: 잭슨 측의 한 변호사는 유@@ 효@@ 한 계약@@ 은 없@@ 었으며 돈@@ 은 대@@ 가@@ 성 없이 주@@ 어진 것이라고 말했습니다 . INFO: TARGET: a lawyer for Jackson says there was never a v@@ ali@@ d agreement , and that the money was given fre@@ ely . INFO: SAMPLE: what was not free , but he says to pr@@ os@@ ful citing something they have enti@@ Senator child . INFO: SOURCE: 총격@@ 범@@ 과 그의 여자@@ 친구 사이 의 말@@ 다@@ 툼 때문에 이번 공격이 발생@@ 했다는 , 확인@@ 되지 않은 보도가 있었다 . INFO: TARGET: un@@ confirmed reports say the attack followed an ar@@ gu@@ ment between the gunman and his girlfriend . INFO: SAMPLE: police had left the militants with his wife , cos@@ the@@ gh@@ ics and two were monitoring . INFO: SOURCE: 왕@@ 복@@ 선은 이틀@@ 간 우주 정거@@ 장을 조사한 뒤 호주 남@@ 동부 3@@ 36@@ km 위에서 작업을 종@@ 료@@ 할 예정이다 . INFO: TARGET: the shuttle 's two @-@ day ch@@ as@@ e of the space station ended about 2@@ 10 miles above southeastern Australia . INFO: SAMPLE: captain R@@ on@@ submitted Pat@@ g@@ el will were vi@@ fi@@ ory , as Ch@@ ang@@ am@@ land international space station will maintain to Tibet . Be@@ at@@ ari televised congratul@@ ate the space station INFO: SOURCE: 한국은 또한 전통적인 경제 논@@ 리에 명@@ 백@@ 히 거@@ 슬@@ 리는 이런 변화를 이@@ 룩@@ 해@@ 내@@ 길 바라고 있다 . INFO: TARGET: Korea also hopes to accompl@@ ish the shift in apparent de@@ fi@@ ance of con@@ v@@ enti@@ onal economic lo@@ gi@@ c . INFO: SAMPLE: Korea also ay which for so balance in@@ land . INFO: SOURCE: 크레@@ 디는 " 체포 당시 15@@ 달 된 아들을 안@@ 고 있었다 " 고 말했다 . INFO: TARGET: " I was holding my baby at the time , " said McC@@ ready of her 15 @-@ month @-@ old son . INFO: SAMPLE: ing walked him in London , " P@@ u@@ k@@ crowd said . INFO: SOURCE: 이슬람 반군@@ 과의 이번 분쟁@@ 은 199@@ 0@@ 년 내@@ 전 이후 레바논@@ 에서 발생한 최악의 유혈@@ 충돌로 기록@@ 되고 있다 . INFO: TARGET: the sometimes @-@ fi@@ er@@ ce battles mark the worst internal violence since the end of Lebanon 's civil war in 1990 . INFO: SAMPLE: Su@@ am@@ there still show Robert A@@ March in the biggest looks at northern Iraq in reasons every year for political fighting momentum ; INFO: SOURCE: T@@ M@@ Z@@ 닷@@ 컴@@ 은 듀@@ 브@@ 로가 네@@ 버@@ 더@@ 주 라@@ 스@@ 베이@@ 거@@ 스에 있는 자택에서 사망했다고 보도했다 . INFO: TARGET: Du@@ B@@ row died at his home in Las Vegas , Nevada , according to T@@ M@@ Z@@ .@@ com@@ . INFO: SAMPLE 0: the A@@ T@@ A said he was killed in a hospital in the Los Angeles capital . Cost/Len/Avg -34.647682189941406/76/-0.45589055513080795 INFO: SAMPLE 1: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles capital . Cost/Len/Avg -34.67599105834961/80/-0.43344988822937014 INFO: SAMPLE 2: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles area . Cost/Len/Avg -34.80378341674805/77/-0.4519971872304941 INFO: SAMPLE 3: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of C@@ as@@ e . Cost/Len/Avg -43.47795104980469/91/-0.4777796818659856 INFO: SAMPLE 4: the A@@ T@@ A said he was killed in a hospital in the Los Angeles city of C@@ as@@ e . Cost/Len/Avg -43.573909759521484/87/-0.5008495374657642 INFO: SAMPLE 5: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of B@@ ri@@ ot@@ a . Cost/Len/Avg -45.60481262207031/96/-0.4750501314798991 INFO: SAMPLE 6: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of P@@ h@@ ol@@ o . Cost/Len/Avg -45.89029312133789/95/-0.48305571706671463 INFO: SAMPLE 7: the A@@ T@@ A said he was killed in a hospital in the Los Angeles city of P@@ h@@ ol@@ o . Cost/Len/Avg -45.95040512084961/91/-0.5049495068225232 INFO: SAMPLE 8: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of P@@ h@@ il@@ d . Cost/Len/Avg -46.11641311645508/95/-0.4854359275416324 INFO: SAMPLE 9: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of B@@ ri@@ ot@@ a , the Associated Press reported . Cost/Len/Avg -50.92622756958008/128/-0.39786115288734436 INFO: SAMPLE 10: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of B@@ ri@@ ot@@ a , the AP reported . Cost/Len/Avg -51.131935119628906/114/-0.44852574666341144 INFO: SAMPLE 11: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of B@@ ri@@ ot@@ a , the official said . Cost/Len/Avg -51.474334716796875/116/-0.4437442647999731 INFO: SOURCE: 추@@ 후 에 요@@ 금이 적용@@ 될 수도 있지만 현@@ 재@@ 는 무@@ 료@@ 다 . INFO: TARGET: initially , there will be no f@@ ee to apply for travel author@@ iz@@ ation , although one may be imposed later . INFO: SAMPLE 0: but it is expected to be expected to be held . Cost/Len/Avg -23.384572982788086/47/-0.4975441060167678 INFO: SAMPLE 1: but it is expected to be expected to be released . Cost/Len/Avg -23.43704605102539/51/-0.4595499225691253 INFO: SAMPLE 2: but it is expected to be expected to be held , but it 's not going to be . Cost/Len/Avg -37.967037200927734/80/-0.47458796501159667 INFO: SAMPLE 3: but it is expected to be expected to be held , but it 's not enough to be . Cost/Len/Avg -38.5172119140625/81/-0.47552113474151236 INFO: SAMPLE 4: but it is expected to be expected to be held , but it 's not going to be able . Cost/Len/Avg -40.010398864746094/85/-0.4707105748793658 INFO: SAMPLE 5: but it is expected to be expected to be held , but it will not be able to be . Cost/Len/Avg -40.27976989746094/79/-0.5098705050311511 INFO: SAMPLE 6: but it is expected to be expected to be held , but it will not be able to be able . Cost/Len/Avg -42.25888442993164/84/-0.5030819574991862 INFO: SAMPLE 7: but it is expected to be expected to be held , but it 's not going to be able to be . Cost/Len/Avg -43.747947692871094/91/-0.4807466779436384 INFO: SAMPLE 8: but it is expected to be expected to be held , but it 's not going to be able to go . Cost/Len/Avg -44.41551208496094/91/-0.48808255038418613 INFO: SAMPLE 9: but it is expected to be expected to be held , but it will not be able to be able to be . Cost/Len/Avg -46.169952392578125/90/-0.5129994710286458 INFO: SAMPLE 10: but it is expected to be expected to be held , but it 's not going to be able to be able . Cost/Len/Avg -46.30955505371094/96/-0.4823911984761556 INFO: SAMPLE 11: but it is expected to be expected to be held , but it will not be able to be able to be able . Cost/Len/Avg -48.63726806640625/95/-0.5119712428042763 INFO: SOURCE: 일부 펀@@ 드@@ 매@@ 니@@ 저는 " 기업 이@@ 윤@@ 이 증가@@ 해 주식@@ 시장이 활@@ 성@@ 화@@ 됐다 " 고 전했다 . INFO: TARGET: fund man@@ ag@@ ers said strong corporate profit growth was supporting stocks . INFO: SAMPLE 0: according to the company , the company said , " it 's a lot of growth . Cost/Len/Avg -37.93861770629883/82/-0.4626660695890101 INFO: SAMPLE 1: according to the company , the company said , " it 's a lot of recession . Cost/Len/Avg -37.9777946472168/85/-0.4467975840849035 INFO: SAMPLE 2: according to the company , the company said , " it 's a lot of economic growth . Cost/Len/Avg -38.13514709472656/91/-0.4190675504915007 INFO: SAMPLE 3: according to the company , the company said , " it 's a lot of interest rates . Cost/Len/Avg -38.815006256103516/90/-0.43127784729003904 INFO: SAMPLE 4: according to the company , the company said , " it 's a lot of recession . " Cost/Len/Avg -38.95130157470703/92/-0.4233837127685547 INFO: SAMPLE 5: according to the company , the company said , " it 's a lot of economic growth . " Cost/Len/Avg -39.129940032958984/98/-0.3992851023771325 INFO: SAMPLE 6: according to the company , the company said , " it 's a lot of interest rates . " Cost/Len/Avg -39.95598602294922/97/-0.41191738167988884 INFO: SAMPLE 7: according to the company , the company said , " it 's a lot of recession , " he said . Cost/Len/Avg -40.88528060913086/102/-0.4008360844032437 INFO: SAMPLE 8: according to the company , the company said , " it 's a lot of economic growth , " he said . Cost/Len/Avg -41.4897346496582/108/-0.38416420971905746 INFO: SAMPLE 9: according to the company , the company said , " it 's a lot of interest rates , " he said . Cost/Len/Avg -41.738468170166016/107/-0.3900791417772525 INFO: SAMPLE 10: according to the company , the company said , " it 's a lot of the economy , " he said . Cost/Len/Avg -42.56211471557617/104/-0.4092511030343863 INFO: SAMPLE 11: according to the company , the company said , " it 's a lot of economic growth , " the company said . Cost/Len/Avg -43.87964630126953/117/-0.37503971197666264 INFO: SOURCE: 이 바이러@@ 스는 서@@ 쪽으로 빠르게 퍼@@ 지고 있으며 , 관리들은 이것이 곧 서부 해안@@ 에@@ 까지 이를 것으로 우려하고 있다 . INFO: TARGET: the virus has been spreading quickly w@@ est@@ ward , with officials f@@ ear@@ ing it will soon reach the West Coast . INFO: SAMPLE 0: the virus is expected to be in the region . Cost/Len/Avg -21.322120666503906/44/-0.4845936515114524 INFO: SAMPLE 1: the virus is expected to be in the region , officials said . Cost/Len/Avg -23.85358238220215/61/-0.39104233413446143 INFO: SAMPLE 2: the virus is expected to be in the area , officials said . Cost/Len/Avg -24.124916076660156/59/-0.40889688265525687 INFO: SAMPLE 3: the virus is expected to be in the region , but officials say they are still in the region . Cost/Len/Avg -40.866912841796875/93/-0.4394291703419019 INFO: SAMPLE 4: the virus is expected to be in the region , but officials said it would be in the region . Cost/Len/Avg -41.33380126953125/91/-0.45421759636847525 INFO: SAMPLE 5: the virus is expected to be in the region , but officials say they are still in the area . Cost/Len/Avg -41.385223388671875/91/-0.45478267460078986 INFO: SAMPLE 6: the virus is expected to be in the region , but officials say they are still in the border . Cost/Len/Avg -41.601173400878906/93/-0.44732444517074094 INFO: SAMPLE 7: the virus is expected to be in the region , but officials said it would be in the area . Cost/Len/Avg -41.86735916137695/89/-0.470419765858168 INFO: SAMPLE 8: the virus is expected to be in the region , but officials say they are still in the town . Cost/Len/Avg -41.90886306762695/91/-0.4605369567871094 INFO: SAMPLE 9: the virus is expected to be in the region , but officials said it would be in the border . Cost/Len/Avg -41.95485305786133/91/-0.4610423412951794 INFO: SAMPLE 10: the virus is expected to be in the region , but officials say they are still in the city . Cost/Len/Avg -42.01584243774414/91/-0.4617125542609246 INFO: SAMPLE 11: the virus is expected to be in the region , but officials say they are still in the town of the region . Cost/Len/Avg -46.920806884765625/105/-0.44686482747395834 INFO: SOURCE: 잭슨 측의 한 변호사는 유@@ 효@@ 한 계약@@ 은 없@@ 었으며 돈@@ 은 대@@ 가@@ 성 없이 주@@ 어진 것이라고 말했습니다 . INFO: TARGET: a lawyer for Jackson says there was never a v@@ ali@@ d agreement , and that the money was given fre@@ ely . INFO: SAMPLE 0: the judge says it 's not clear that it 's not clear that it 's not enough . Cost/Len/Avg -35.635536193847656/91/-0.3915992988334907 INFO: SAMPLE 1: the judge says it 's not clear that it 's not clear that it 's not enough to be . Cost/Len/Avg -39.26071548461914/97/-0.40474964417133136 INFO: SAMPLE 2: the judge says it 's not clear that it was not clear that it 's not enough to be . Cost/Len/Avg -39.52196502685547/93/-0.4249673658801663 INFO: SAMPLE 3: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not enough . Cost/Len/Avg -45.02193832397461/117/-0.38480289165790266 INFO: SAMPLE 4: the judge says it 's not clear that it was not clear that it 's not clear that it 's not enough . Cost/Len/Avg -45.23143768310547/113/-0.40027820958500415 INFO: SAMPLE 5: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not a lot . Cost/Len/Avg -47.7430419921875/116/-0.41157794820851296 INFO: SAMPLE 6: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not enough to be . Cost/Len/Avg -48.963233947753906/123/-0.3980750727459667 INFO: SAMPLE 7: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not a little bit . Cost/Len/Avg -49.09718704223633/123/-0.3991641222946043 INFO: SAMPLE 8: the judge says it 's not clear that it was not clear that it 's not clear that it 's not enough to be . Cost/Len/Avg -49.15306091308594/119/-0.41305093204273896 INFO: SAMPLE 9: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not enough to be . " Cost/Len/Avg -52.79254150390625/130/-0.40609647310697117 INFO: SAMPLE 10: the judge says it 's not clear that it was not clear that it 's not clear that it 's not enough to be . " Cost/Len/Avg -53.00259017944336/126/-0.42065547761462985 INFO: SAMPLE 11: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not a little bit . " Cost/Len/Avg -53.03470993041992/130/-0.40795930715707635 INFO: SOURCE: 총격@@ 범@@ 과 그의 여자@@ 친구 사이 의 말@@ 다@@ 툼 때문에 이번 공격이 발생@@ 했다는 , 확인@@ 되지 않은 보도가 있었다 . INFO: TARGET: un@@ confirmed reports say the attack followed an ar@@ gu@@ ment between the gunman and his girlfriend . INFO: SAMPLE 0: the incident was found in the attack . Cost/Len/Avg -16.111059188842773/39/-0.4131040817651993 INFO: SAMPLE 1: the incident was found in the attack , but it was not reported . Cost/Len/Avg -25.738344192504883/65/-0.39597452603853667 INFO: SAMPLE 2: the incident was found in the attack , but it was not immediately reported . Cost/Len/Avg -26.580915451049805/77/-0.34520669416947797 INFO: SAMPLE 3: the incident was found in the attack , but it was not immediately known . Cost/Len/Avg -27.21120262145996/74/-0.36771895434405355 INFO: SAMPLE 4: the incident was found in the attack , but the incident was not immediately known . Cost/Len/Avg -28.442596435546875/84/-0.3386023385184152 INFO: SAMPLE 5: the incident was found in the attack , but it was no evidence of the attack . Cost/Len/Avg -29.79374885559082/78/-0.3819711391742413 INFO: SAMPLE 6: the incident was found in the attack , but it was not immediately known as the attack . Cost/Len/Avg -32.138160705566406/88/-0.3652063716541637 INFO: SAMPLE 7: the incident was found in the attack , but it was not clear that the attack was not known . Cost/Len/Avg -34.64654541015625/92/-0.37659288489300274 INFO: SAMPLE 8: the incident was found in the attack , but it was not clear that the incident was not known . Cost/Len/Avg -34.85503005981445/94/-0.37079819212568566 INFO: SAMPLE 9: the incident was found in the attack , but it was not immediately clear that he was not known . Cost/Len/Avg -35.30630874633789/96/-0.3677740494410197 INFO: SAMPLE 10: the incident was found in the attack , but it was not immediately clear that the attack was not reported . Cost/Len/Avg -35.64750289916992/107/-0.33315423270252265 INFO: SAMPLE 11: the incident was found in the attack , but it was not clear that the attack was not immediately known . Cost/Len/Avg -35.765377044677734/104/-0.34389785619882435 INFO: SOURCE: 왕@@ 복@@ 선은 이틀@@ 간 우주 정거@@ 장을 조사한 뒤 호주 남@@ 동부 3@@ 36@@ km 위에서 작업을 종@@ 료@@ 할 예정이다 . INFO: TARGET: the shuttle 's two @-@ day ch@@ as@@ e of the space station ended about 2@@ 10 miles above southeastern Australia . INFO: SAMPLE 0: London , England ( CNN ) Prince Harry is expected to be in the first time after the first time of the space station after the first time of the space station . Cost/Len/Avg -52.85684585571289/160/-0.33035528659820557 /home/amitkumar.rs.cse17.iitbhu/nematus/nematus/train.py:290: RuntimeWarning: divide by zero encountered in true_divide i, sample, cost, len(sample), cost/len(sample)) INFO: SAMPLE 1: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 2: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 3: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 4: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 5: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 6: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 7: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 8: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 9: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 10: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 11: Cost/Len/Avg -10000000.0/0/-inf INFO: SOURCE: 한국은 또한 전통적인 경제 논@@ 리에 명@@ 백@@ 히 거@@ 슬@@ 리는 이런 변화를 이@@ 룩@@ 해@@ 내@@ 길 바라고 있다 . INFO: TARGET: Korea also hopes to accompl@@ ish the shift in apparent de@@ fi@@ ance of con@@ v@@ enti@@ onal economic lo@@ gi@@ c . INFO: SAMPLE 0: South Korea also said it will also be a new economic crisis . Cost/Len/Avg -24.42115020751953/62/-0.3938895194761215 INFO: SAMPLE 1: South Korea also said that Korea is also likely to be a new economic crisis . Cost/Len/Avg -29.65618896484375/78/-0.3802075508313301 INFO: SAMPLE 2: South Korea also said that Korea is also likely to be a new economy . Cost/Len/Avg -29.85259246826172/70/-0.4264656066894531 INFO: SAMPLE 3: South Korea also said that Korea is also likely to make a new economic crisis . Cost/Len/Avg -29.87238121032715/80/-0.3734047651290894 INFO: SAMPLE 4: South Korea also said that Korea is also likely to be a major economic crisis . Cost/Len/Avg -29.930070877075195/80/-0.37412588596343993 INFO: SAMPLE 5: South Korea also said that Korea is also likely to take a new economic crisis . Cost/Len/Avg -30.075361251831055/80/-0.3759420156478882 INFO: SAMPLE 6: South Korea also said that Korea is also likely to get a new economic crisis . Cost/Len/Avg -30.59653091430664/79/-0.3872978596747676 INFO: SAMPLE 7: South Korea also said that Korea is also likely to have a new economic crisis . Cost/Len/Avg -30.841949462890625/80/-0.3855243682861328 INFO: SAMPLE 8: South Korea also said that Korea is also likely to be a new economic crisis in the world . Cost/Len/Avg -35.9841194152832/91/-0.3954298836844308 INFO: SAMPLE 9: South Korea also said that Korea is also likely to make a new economic crisis in the world . Cost/Len/Avg -36.09120178222656/93/-0.3880774385185652 INFO: SAMPLE 10: South Korea also said that Korea is also likely to be a new economic crisis in the country . Cost/Len/Avg -37.242855072021484/93/-0.40046080722603744 INFO: SAMPLE 11: South Korea also said that Korea is also likely to make a new economic crisis in the country . Cost/Len/Avg -37.32313919067383/95/-0.392875149375514 INFO: SOURCE: 크레@@ 디는 " 체포 당시 15@@ 달 된 아들을 안@@ 고 있었다 " 고 말했다 . INFO: TARGET: " I was holding my baby at the time , " said McC@@ ready of her 15 @-@ month @-@ old son . INFO: SAMPLE 0: he said he had been arrested last month . Cost/Len/Avg -18.11133575439453/42/-0.43122227986653644 INFO: SAMPLE 1: he said he had been arrested last year . Cost/Len/Avg -18.825273513793945/41/-0.45915301253155966 INFO: SAMPLE 2: he said he was taken to his wife . Cost/Len/Avg -19.07192039489746/35/-0.5449120112827845 INFO: SAMPLE 3: he said he had been arrested last month , " he said . Cost/Len/Avg -22.439355850219727/59/-0.3803280652579615 INFO: SAMPLE 4: he said he was taken to his wife , " he said . Cost/Len/Avg -22.90827751159668/52/-0.4405437982999362 INFO: SAMPLE 5: he said he had been arrested last month , " she said . Cost/Len/Avg -23.00881576538086/60/-0.38348026275634767 INFO: SAMPLE 6: he said he was taken to his wife , " she said . Cost/Len/Avg -23.43367576599121/53/-0.44214482577341907 INFO: SAMPLE 7: he said he had been arrested last month , " he said . " Cost/Len/Avg -28.204511642456055/66/-0.4273410854917584 INFO: SAMPLE 8: he said he was taken to his wife , " he said . " Cost/Len/Avg -28.393362045288086/59/-0.4812434244964082 INFO: SAMPLE 9: he said he had been arrested last month , " he said . " it was a child . Cost/Len/Avg -36.910091400146484/83/-0.44469989638730706 INFO: SAMPLE 10: he said he had been arrested last month , " he said . " it was a child . " Cost/Len/Avg -38.843902587890625/90/-0.43159891764322916 INFO: SAMPLE 11: he said he had been arrested last month , " he said . " it was not a child . Cost/Len/Avg -39.1064338684082/87/-0.44949923986676094 INFO: SOURCE: 이슬람 반군@@ 과의 이번 분쟁@@ 은 199@@ 0@@ 년 내@@ 전 이후 레바논@@ 에서 발생한 최악의 유혈@@ 충돌로 기록@@ 되고 있다 . INFO: TARGET: the sometimes @-@ fi@@ er@@ ce battles mark the worst internal violence since the end of Lebanon 's civil war in 1990 . INFO: SAMPLE 0: the Islamic militant group has been held in the northern city of southern Lebanon . Cost/Len/Avg -30.197078704833984/84/-0.35948903220040457 INFO: SAMPLE 1: the Islamic militant group has been held in the northern city of northern Lebanon . Cost/Len/Avg -30.308401107788086/84/-0.3608142989022391 INFO: SAMPLE 2: the Islamic militant group has been held in the northern city of northern Iraq . Cost/Len/Avg -30.530426025390625/81/-0.37691883981963736 INFO: SAMPLE 3: the Islamic militant group has been held in the northern city of northern Iraq in the region . Cost/Len/Avg -37.04274368286133/95/-0.38992361771432976 INFO: SAMPLE 4: the Islamic militant group has been held in the northern city of northern Lebanon , which has been in the region . Cost/Len/Avg -44.6102180480957/115/-0.3879149395486583 INFO: SAMPLE 5: the Islamic militant group has been held in the northern city of southern Lebanon , which has been in the region . Cost/Len/Avg -44.716522216796875/115/-0.38883932362432067 INFO: SAMPLE 6: the Islamic militant group has been held in the northern city of northern Iraq , which has been in the region . Cost/Len/Avg -44.74925994873047/112/-0.39954696382795063 INFO: SAMPLE 7: the Islamic militant group has been held in the northern city of northern Lebanon , which has been the worst in the region . Cost/Len/Avg -48.818878173828125/125/-0.390551025390625 INFO: SAMPLE 8: the Islamic militant group has been held in the northern city of southern Lebanon , which has been the worst in the region . Cost/Len/Avg -48.91313171386719/125/-0.3913050537109375 INFO: SAMPLE 9: the Islamic militant group has been held in the northern city of northern Iraq , which has been the worst in the region . Cost/Len/Avg -48.95859909057617/122/-0.4012999925457063 INFO: SAMPLE 10: the Islamic militant group has been held in the northern city of northern Lebanon , which has been the worst in the country . Cost/Len/Avg -49.13996124267578/126/-0.38999969240218874 INFO: SAMPLE 11: the Islamic militant group has been held in the northern city of southern Lebanon , which has been the worst in the country . Cost/Len/Avg -49.23438262939453/126/-0.3907490684872582 INFO: Seen 8 INFO: Seen 20 INFO: Seen 33 INFO: Seen 48 INFO: Seen 63 INFO: Seen 79 INFO: Seen 96 INFO: Seen 114 INFO: Seen 132 INFO: Seen 151 INFO: Seen 171 INFO: Seen 191 INFO: Seen 212 INFO: Seen 234 INFO: Seen 256 INFO: Seen 279 INFO: Seen 304 INFO: Seen 329 INFO: Seen 355 INFO: Seen 382 INFO: Seen 410 INFO: Seen 439 INFO: Seen 470 INFO: Seen 502 INFO: Seen 535 INFO: Seen 570 INFO: Seen 608 INFO: Seen 649 INFO: Seen 694 INFO: Seen 746 INFO: Seen 800 INFO: Seen 811 INFO: Seen 828 INFO: Seen 848 INFO: Seen 871 INFO: Seen 898 INFO: Seen 930 INFO: Seen 970 INFO: Seen 1000 INFO: Validation cross entropy (AVG/SUM/N_SENTS/N_TOKENS): 187.86940924072266 187869.40924072266 1000 33910 INFO: scripts/..//model/tmpv6367j94/model is not in all_model_checkpoint_paths. Manually adding it. INFO: Starting external validation. INFO: NOTE: Length of translations is capped to 200 INFO: Translated 40 sents INFO: Translated 80 sents INFO: Translated 120 sents INFO: Translated 160 sents INFO: Translated 200 sents INFO: Translated 240 sents INFO: Translated 280 sents INFO: Translated 320 sents INFO: Translated 360 sents INFO: Translated 400 sents INFO: Translated 440 sents INFO: Translated 480 sents INFO: Translated 520 sents INFO: Translated 560 sents INFO: Translated 600 sents INFO: Translated 640 sents INFO: Translated 680 sents INFO: Translated 720 sents INFO: Translated 760 sents INFO: Translated 800 sents INFO: Translated 840 sents INFO: Translated 880 sents INFO: Translated 920 sents INFO: Translated 960 sents INFO: Translated 1000 sents INFO: Translated 1000 sents in 23.049481868743896 sec. Speed 43.38492317070448 sents/sec Traceback (most recent call last): File "/home/amitkumar.rs.cse17.iitbhu/nematus/nematus/train.py", line 522, in train(config, sess) File "/home/amitkumar.rs.cse17.iitbhu/nematus/nematus/train.py", line 322, in train score = validate_with_script(sess, beam_search_sampler) File "/home/amitkumar.rs.cse17.iitbhu/nematus/nematus/train.py", line 426, in validate_with_script stderr=subprocess.PIPE) File "/opt/ohpc/pub/apps/anaconda3/envs/py35/lib/python3.5/subprocess.py", line 676, in init restore_signals, start_new_session) File "/opt/ohpc/pub/apps/anaconda3/envs/py35/lib/python3.5/subprocess.py", line 1289, in _execute_child raise child_exception_type(errno_num, err_msg) PermissionError: [Errno 13] Permission denied
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/EdinburghNLP/nematus/issues/120, or unsubscribe https://github.com/notifications/unsubscribe-auth/AADCXSZQ6E2TWX72AUKWGVLTGVKQJANCNFSM42IKDOUA.
Here is the training file
!/bin/sh
script_dir=
dirname $0main_dir=$script_dir/../ data_dir=$main_dir/data working_dir=$main_dir/modeldevice=0,1
independent variables (toolkit locations)
. $main_dir/../vars
language-dependent variables (source and target language)
. $main_dir/vars
CUDA_VISIBLE_DEVICES=$device python $nematus_home/nematus/train.py \ --model $working_dir/model \ --datasets $data_dir/korean-english-park.train.bpe.$src $data_dir/korean-english-park.train.bpe.$trg \ --valid_datasets $data_dir/korean-english-park.dev.bpe.$src $data_dir/korean-english-park.dev.bpe.$trg \ --dictionaries $data_dir/korean-english-park.train.bpe.ko.json $data_dir/korean-english-park.train.bpe.en.json \ --valid_script $script_dir/validate.sh \ --reload latest_checkpoint \ --dim_word 512 \ --dim 1024 \ --lrate 0.0005 \ --optimizer adam \ --maxlen 200 \ --batch_size 80 \ --valid_batch_size 40 \ --validFreq 10000 \ --dispFreq 1000 \ --saveFreq 30000 \ --sampleFreq 10000 \ --tie_decoder_embeddings \ --layer_normalisation \ --dec_base_recurrence_transition_depth 2 \ --enc_recurrence_transition_depth 2 \ --rnn_dropout_hidden 0.5 \ --rnn_dropout_embedding 0.5 \ --rnn_dropout_source 0.3 \ --rnn_dropout_target 0.3 \ --label_smoothing 0.2 \ --token_batch_size 1000 \ --valid_token_batch_size 1000 \ --patience 10
validate.sh file mentioned below:
!/bin/sh
script_dir=
dirname $0main_dir=$script_dir/../ data_dir=$main_dir/data working_dir=$main_dir/modeldevice=0,1
language-dependent variables (source and target language)
. $main_dir/vars
dev_prefix=korean-english-park.dev dev=$dev_prefix.bpe.$src ref=$dev_prefix.$trg prefix=$working_dir/model
CUDA_VISIBLE_DEVICES=$device python $nematus_home/nematus/translate.py \ -m $prefix \ -i $data_dir/$dev \ -o $working_dir/$dev.output.dev \ -k 5 \ -n
bash $script_dir/postprocess.sh < $working_dir/$dev.output.dev > $working_dir/$dev.output.postprocessed.dev
BEST=
cat ${prefix}_best_bleu || echo 0$nematus_home/data/multi-bleu-detok.perl /scratch/amitkumar.rs.cse17.iitbhu/ko-en/parallel/$ref < $working_dir/$dev.output.postprocessed.dev >> ${prefix}_bleu_scores BLEU=$nematus_home/data/multi-bleu-detok.perl /scratch/amitkumar.rs.cse17.iitbhu/ko-en/parallel/$ref < $working_dir/$dev.output.postprocessed.dev | cut -f 3 -d ' ' | cut -f 1 -d ','BETTER=echo "$BLEU > $BEST" | bcecho "BLEU = $BLEU"
echo "BLEU = $BLEU"
if [ "$BETTER" = "1" ]; then echo "new best; saving" echo $BLEU > ${prefix}_best_bleu cp ${prefix}.json ${prefix}.best_bleu.json fi
Following errors in training:
2021-04-02 07:11:20.787601: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 INFO: Namespace(adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08, batch_size=80, beam_freq=10000, beam_size=12, clip_c=1.0, datasets=['scripts/..//data/korean-english-park.train.bpe.ko', 'scripts/..//data/korean-english-park.train.bpe.en'], decay_c=0.0, dictionaries=['scripts/..//data/korean-english-park.train.bpe.ko.json', 'scripts/..//data/korean-english-park.train.bpe.en.json'], dim_per_factor=[512], disp_freq=1000, embedding_size=512, exponential_smoothing=0.0, factors=1, finish_after=10000000, gradient_aggregation_steps=1, keep_train_set_in_memory=False, label_smoothing=0.2, layer_normalization_type='layernorm', learning_rate=0.0005, learning_schedule='constant', loss_function='cross-entropy', map_decay_c=0.0, max_epochs=5000, max_len_a=1.5, max_len_b=5, max_sentences_of_sampling=0, max_sentences_per_device=0, max_tokens_per_device=0, maxibatch_size=20, maxlen=200, model_type='rnn', model_version=0.2, mrt_alpha=0.005, mrt_loss='SENTENCEBLEU n=4', mrt_ml_mix=0, mrt_reference=False, n_best=False, normalization_alpha=0.0, optimizer='adam', output_hidden_activation='tanh', patience=10, plateau_steps=0, preprocess_script=None, print_per_token_pro=False, prior_model=None, reload='latest_checkpoint', reload_training_progress=True, rnn_dec_base_transition_depth=2, rnn_dec_deep_context=False, rnn_dec_depth=1, rnn_dec_high_transition_depth=1, rnn_dropout_embedding=0.5, rnn_dropout_hidden=0.5, rnn_dropout_source=0.3, rnn_dropout_target=0.3, rnn_enc_depth=1, rnn_enc_transition_depth=2, rnn_layer_normalization=True, rnn_lexical_model=False, rnn_use_dropout=True, sample_freq=10000, sample_way='beam_search', samplesN=100, sampling_temperature=1.0, save_freq=30000, saveto='scripts/..//model/model', shuffle_each_epoch=True, softmax_mixture_size=1, sort_by_length=True, source_dataset='scripts/..//data/korean-english-park.train.bpe.ko', source_dicts=['scripts/..//data/korean-english-park.train.bpe.ko.json'], source_vocab_sizes=[9910], state_size=1024, summary_dir=None, summary_freq=0, target_dataset='scripts/..//data/korean-english-park.train.bpe.en', target_dict='scripts/..//data/korean-english-park.train.bpe.en.json', target_embedding_size=512, target_vocab_size=5953, theano_compat=False, tie_decoder_embeddings=True, tie_encoder_decoder_embeddings=False, token_batch_size=1000, transformer_dec_depth=6, transformer_drophead=0.0, transformer_dropout_attn=0.1, transformer_dropout_embeddings=0.1, transformer_dropout_relu=0.1, transformer_dropout_residual=0.1, transformer_enc_depth=6, transformer_ffn_hidden_size=2048, transformer_num_heads=8, translation_maxlen=200, translation_strategy='beam_search', valid_batch_size=40, valid_bleu_source_dataset='scripts/..//data/korean-english-park.dev.bpe.ko', valid_datasets=['scripts/..//data/korean-english-park.dev.bpe.ko', 'scripts/..//data/korean-english-park.dev.bpe.en'], valid_freq=10000, valid_script='scripts/validate.sh', valid_source_dataset='scripts/..//data/korean-english-park.dev.bpe.ko', valid_target_dataset='scripts/..//data/korean-english-park.dev.bpe.en', valid_token_batch_size=1000, warmup_steps=8000) 2021-04-02 07:11:41.096118: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN)to use the following CPU instructions in performance-critical operations: AVX2 AVX512F FMA To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags. 2021-04-02 07:11:41.108886: I tensorflow/core/platform/profile_utils/cpu_utils.cc:104] CPU Frequency: 2400000000 Hz 2021-04-02 07:11:41.110970: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x5612befedba0 initialized for platform Host (this does not guarantee that XLA will be used). Devices: 2021-04-02 07:11:41.111051: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version 2021-04-02 07:11:41.116171: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcuda.so.1 2021-04-02 07:11:42.267591: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x5612bf0af6e0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices: 2021-04-02 07:11:42.267741: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Tesla V100-PCIE-16GB, Compute Capability 7.0 2021-04-02 07:11:42.270571: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1716] Found device 0 with properties: pciBusID: 0000:5e:00.0 name: Tesla V100-PCIE-16GB computeCapability: 7.0 coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 15.75GiB deviceMemoryBandwidth: 836.37GiB/s 2021-04-02 07:11:42.270674: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 2021-04-02 07:11:42.276043: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10 2021-04-02 07:11:42.279407: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcufft.so.10 2021-04-02 07:11:42.280411: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcurand.so.10 2021-04-02 07:11:42.283420: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusolver.so.10 2021-04-02 07:11:42.285053: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusparse.so.10 2021-04-02 07:11:47.565824: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudnn.so.7 2021-04-02 07:11:47.571036: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1858] Adding visible gpu devices: 0 2021-04-02 07:11:47.571172: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 2021-04-02 07:11:49.279069: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1257] Device interconnect StreamExecutor with strength 1 edge matrix: 2021-04-02 07:11:49.279154: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1263] 0 2021-04-02 07:11:49.279172: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1276] 0: N 2021-04-02 07:11:49.281268: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1402] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 14729 MB memory) -> physical GPU (device: 0, name: Tesla V100-PCIE-16GB, pci bus id: 0000:5e:00.0, compute capability: 7.0) 2021-04-02 07:11:49.286117: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1716] Found device 0 with properties: pciBusID: 0000:5e:00.0 name: Tesla V100-PCIE-16GB computeCapability: 7.0 coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 15.75GiB deviceMemoryBandwidth: 836.37GiB/s 2021-04-02 07:11:49.286209: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 2021-04-02 07:11:49.286244: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10 2021-04-02 07:11:49.286269: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcufft.so.10 2021-04-02 07:11:49.286293: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcurand.so.10 2021-04-02 07:11:49.286316: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusolver.so.10 2021-04-02 07:11:49.286340: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusparse.so.10 2021-04-02 07:11:49.286364: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudnn.so.7 2021-04-02 07:11:49.288785: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1858] Adding visible gpu devices: 0 2021-04-02 07:11:49.288839: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1257] Device interconnect StreamExecutor with strength 1 edge matrix: 2021-04-02 07:11:49.288879: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1263] 0 2021-04-02 07:11:49.288899: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1276] 0: N 2021-04-02 07:11:49.291537: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1402] Created TensorFlow device (/device:GPU:0 with 14729 MB memory) -> physical GPU (device: 0, name: Tesla V100-PCIE-16GB, pci bus id: 0000:5e:00.0, compute capability: 7.0) INFO: Building model... WARNING: From /home/amitkumar.rs.cse17.iitbhu/nematus/nematus/rnn_model.py:59: dropout (from tensorflow.python.keras.legacy_tf_layers.core) is deprecated and will be removed in a future version. Instructions for updating: Use keras.layers.dropout instead. WARNING: From /home/amitkumar.rs.cse17.iitbhu/envNem1/lib/python3.5/site-packages/tensorflow-2.3.1-py3.5-linux-x86_64.egg/tensorflow/python/keras/legacy_tf_layers/core.py:271: Layer.apply (from tensorflow.python.keras.engine.base_layer_v1) is deprecated and will be removed in a future version. Instructions for updating: Please use
train(config, sess)
File "/home/amitkumar.rs.cse17.iitbhu/nematus/nematus/train.py", line 322, in train
score = validate_with_script(sess, beam_search_sampler)
File "/home/amitkumar.rs.cse17.iitbhu/nematus/nematus/train.py", line 426, in validate_with_script
stderr=subprocess.PIPE)
File "/opt/ohpc/pub/apps/anaconda3/envs/py35/lib/python3.5/subprocess.py", line 676, in init
restore_signals, start_new_session)
File "/opt/ohpc/pub/apps/anaconda3/envs/py35/lib/python3.5/subprocess.py", line 1289, in _execute_child
raise child_exception_type(errno_num, err_msg)
PermissionError: [Errno 13] Permission denied
layer.__call__method instead. 2021-04-02 07:11:51.977754: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1716] Found device 0 with properties: pciBusID: 0000:5e:00.0 name: Tesla V100-PCIE-16GB computeCapability: 7.0 coreClock: 1.38GHz coreCount: 80 deviceMemorySize: 15.75GiB deviceMemoryBandwidth: 836.37GiB/s 2021-04-02 07:11:51.977826: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1 2021-04-02 07:11:51.977870: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10 2021-04-02 07:11:51.977889: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcufft.so.10 2021-04-02 07:11:51.977907: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcurand.so.10 2021-04-02 07:11:51.977925: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusolver.so.10 2021-04-02 07:11:51.977942: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusparse.so.10 2021-04-02 07:11:51.977965: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudnn.so.7 2021-04-02 07:11:51.979691: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1858] Adding visible gpu devices: 0 INFO: Initializing model parameters from scratch... INFO: Done INFO: Reading data... INFO: Done INFO: Initial uidx=0 INFO: Starting epoch 0 2021-04-02 07:12:23.860174: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10 INFO: [2021-04-02 07:33:59] Epoch: 0 Update: 1000 Loss/word: 6.966987346043379 Words/sec: 696.856504668986 Sents/sec: 21.713141805022023 INFO: [2021-04-02 07:55:22] Epoch: 0 Update: 2000 Loss/word: 6.580592949380649 Words/sec: 703.13946539418 Sents/sec: 22.02518408136046 INFO: [2021-04-02 08:16:18] Epoch: 0 Update: 3000 Loss/word: 6.428336913131417 Words/sec: 722.7127947165809 Sents/sec: 22.798869988266855 INFO: Starting epoch 1 INFO: [2021-04-02 08:37:23] Epoch: 1 Update: 4000 Loss/word: 6.330259053899345 Words/sec: 716.8921478972981 Sents/sec: 22.429987003584493 INFO: [2021-04-02 08:58:25] Epoch: 1 Update: 5000 Loss/word: 6.2662326674608595 Words/sec: 717.9178678249805 Sents/sec: 22.6803643125988 INFO: [2021-04-02 09:19:13] Epoch: 1 Update: 6000 Loss/word: 6.226474244214676 Words/sec: 726.10896397206 Sents/sec: 22.971330964205684 INFO: Starting epoch 2 INFO: [2021-04-02 09:40:21] Epoch: 2 Update: 7000 Loss/word: 6.17927944352242 Words/sec: 714.946995990232 Sents/sec: 22.401948114913043 INFO: [2021-04-02 10:01:17] Epoch: 2 Update: 8000 Loss/word: 6.147852558704621 Words/sec: 719.9293130745198 Sents/sec: 22.550224623889665 INFO: [2021-04-02 10:22:19] Epoch: 2 Update: 9000 Loss/word: 6.124515647550154 Words/sec: 718.4753338888931 Sents/sec: 22.38133746941181 INFO: Starting epoch 3 INFO: [2021-04-02 10:43:16] Epoch: 3 Update: 10000 Loss/word: 6.093494990647863 Words/sec: 722.7513458365172 Sents/sec: 22.890552982759697 INFO: SOURCE: T@@ M@@ Z@@ 닷@@ 컴@@ 은 듀@@ 브@@ 로가 네@@ 버@@ 더@@ 주 라@@ 스@@ 베이@@ 거@@ 스에 있는 자택에서 사망했다고 보도했다 . INFO: TARGET: Du@@ B@@ row died at his home in Las Vegas , Nevada , according to T@@ M@@ Z@@ .@@ com@@ . INFO: SAMPLE: American G@@ reported C@@ D wasn 't Roger C@@ bad C@@ y@@ d@@ y@@ d , Paris and h@@ ol@@ ly behind your rushed safely today . INFO: SOURCE: 추@@ 후 에 요@@ 금이 적용@@ 될 수도 있지만 현@@ 재@@ 는 무@@ 료@@ 다 . INFO: TARGET: initially , there will be no f@@ ee to apply for travel author@@ iz@@ ation , although one may be imposed later . INFO: SAMPLE: after car@@ c@@ ers often be used pa@@ rent , but a gi@@ gh@@ ty ency has been thought to vote oe @-@ on test ! INFO: SOURCE: 일부 펀@@ 드@@ 매@@ 니@@ 저는 " 기업 이@@ 윤@@ 이 증가@@ 해 주식@@ 시장이 활@@ 성@@ 화@@ 됐다 " 고 전했다 . INFO: TARGET: fund man@@ ag@@ ers said strong corporate profit growth was supporting stocks . INFO: SAMPLE: more than $ 3.@@ 5 bailout attacked C@@ m@@ succeed for more barrels a rise , " 29@@ ,000 yen re@@ str@@ ong@@ s to investors . INFO: SOURCE: 이 바이러@@ 스는 서@@ 쪽으로 빠르게 퍼@@ 지고 있으며 , 관리들은 이것이 곧 서부 해안@@ 에@@ 까지 이를 것으로 우려하고 있다 . INFO: TARGET: the virus has been spreading quickly w@@ est@@ ward , with officials f@@ ear@@ ing it will soon reach the West Coast . INFO: SAMPLE: camps are cleared to sea winds and often @-@ anger gl@@ om@@ ing to Iraq as " coast of Hospital always . documentary loc@@ als politicians northwest the route . term of increasing exposure because INFO: SOURCE: 잭슨 측의 한 변호사는 유@@ 효@@ 한 계약@@ 은 없@@ 었으며 돈@@ 은 대@@ 가@@ 성 없이 주@@ 어진 것이라고 말했습니다 . INFO: TARGET: a lawyer for Jackson says there was never a v@@ ali@@ d agreement , and that the money was given fre@@ ely . INFO: SAMPLE: what was not free , but he says to pr@@ os@@ ful citing something they have enti@@ Senator child . INFO: SOURCE: 총격@@ 범@@ 과 그의 여자@@ 친구 사이 의 말@@ 다@@ 툼 때문에 이번 공격이 발생@@ 했다는 , 확인@@ 되지 않은 보도가 있었다 . INFO: TARGET: un@@ confirmed reports say the attack followed an ar@@ gu@@ ment between the gunman and his girlfriend . INFO: SAMPLE: police had left the militants with his wife , cos@@ the@@ gh@@ ics and two were monitoring . INFO: SOURCE: 왕@@ 복@@ 선은 이틀@@ 간 우주 정거@@ 장을 조사한 뒤 호주 남@@ 동부 3@@ 36@@ km 위에서 작업을 종@@ 료@@ 할 예정이다 . INFO: TARGET: the shuttle 's two @-@ day ch@@ as@@ e of the space station ended about 2@@ 10 miles above southeastern Australia . INFO: SAMPLE: captain R@@ on@@ submitted Pat@@ g@@ el will were vi@@ fi@@ ory , as Ch@@ ang@@ am@@ land international space station will maintain to Tibet . Be@@ at@@ ari televised congratul@@ ate the space station INFO: SOURCE: 한국은 또한 전통적인 경제 논@@ 리에 명@@ 백@@ 히 거@@ 슬@@ 리는 이런 변화를 이@@ 룩@@ 해@@ 내@@ 길 바라고 있다 . INFO: TARGET: Korea also hopes to accompl@@ ish the shift in apparent de@@ fi@@ ance of con@@ v@@ enti@@ onal economic lo@@ gi@@ c . INFO: SAMPLE: Korea also ay which for so balance in@@ land . INFO: SOURCE: 크레@@ 디는 " 체포 당시 15@@ 달 된 아들을 안@@ 고 있었다 " 고 말했다 . INFO: TARGET: " I was holding my baby at the time , " said McC@@ ready of her 15 @-@ month @-@ old son . INFO: SAMPLE: ing walked him in London , " P@@ u@@ k@@ crowd said . INFO: SOURCE: 이슬람 반군@@ 과의 이번 분쟁@@ 은 199@@ 0@@ 년 내@@ 전 이후 레바논@@ 에서 발생한 최악의 유혈@@ 충돌로 기록@@ 되고 있다 . INFO: TARGET: the sometimes @-@ fi@@ er@@ ce battles mark the worst internal violence since the end of Lebanon 's civil war in 1990 . INFO: SAMPLE: Su@@ am@@ there still show Robert A@@ March in the biggest looks at northern Iraq in reasons every year for political fighting momentum ; INFO: SOURCE: T@@ M@@ Z@@ 닷@@ 컴@@ 은 듀@@ 브@@ 로가 네@@ 버@@ 더@@ 주 라@@ 스@@ 베이@@ 거@@ 스에 있는 자택에서 사망했다고 보도했다 . INFO: TARGET: Du@@ B@@ row died at his home in Las Vegas , Nevada , according to T@@ M@@ Z@@ .@@ com@@ . INFO: SAMPLE 0: the A@@ T@@ A said he was killed in a hospital in the Los Angeles capital . Cost/Len/Avg -34.647682189941406/76/-0.45589055513080795 INFO: SAMPLE 1: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles capital . Cost/Len/Avg -34.67599105834961/80/-0.43344988822937014 INFO: SAMPLE 2: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles area . Cost/Len/Avg -34.80378341674805/77/-0.4519971872304941 INFO: SAMPLE 3: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of C@@ as@@ e . Cost/Len/Avg -43.47795104980469/91/-0.4777796818659856 INFO: SAMPLE 4: the A@@ T@@ A said he was killed in a hospital in the Los Angeles city of C@@ as@@ e . Cost/Len/Avg -43.573909759521484/87/-0.5008495374657642 INFO: SAMPLE 5: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of B@@ ri@@ ot@@ a . Cost/Len/Avg -45.60481262207031/96/-0.4750501314798991 INFO: SAMPLE 6: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of P@@ h@@ ol@@ o . Cost/Len/Avg -45.89029312133789/95/-0.48305571706671463 INFO: SAMPLE 7: the A@@ T@@ A said he was killed in a hospital in the Los Angeles city of P@@ h@@ ol@@ o . Cost/Len/Avg -45.95040512084961/91/-0.5049495068225232 INFO: SAMPLE 8: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of P@@ h@@ il@@ d . Cost/Len/Avg -46.11641311645508/95/-0.4854359275416324 INFO: SAMPLE 9: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of B@@ ri@@ ot@@ a , the Associated Press reported . Cost/Len/Avg -50.92622756958008/128/-0.39786115288734436 INFO: SAMPLE 10: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of B@@ ri@@ ot@@ a , the AP reported . Cost/Len/Avg -51.131935119628906/114/-0.44852574666341144 INFO: SAMPLE 11: the A@@ T@@ A was killed in a Los Angeles hospital in the Los Angeles city of B@@ ri@@ ot@@ a , the official said . Cost/Len/Avg -51.474334716796875/116/-0.4437442647999731 INFO: SOURCE: 추@@ 후 에 요@@ 금이 적용@@ 될 수도 있지만 현@@ 재@@ 는 무@@ 료@@ 다 . INFO: TARGET: initially , there will be no f@@ ee to apply for travel author@@ iz@@ ation , although one may be imposed later . INFO: SAMPLE 0: but it is expected to be expected to be held . Cost/Len/Avg -23.384572982788086/47/-0.4975441060167678 INFO: SAMPLE 1: but it is expected to be expected to be released . Cost/Len/Avg -23.43704605102539/51/-0.4595499225691253 INFO: SAMPLE 2: but it is expected to be expected to be held , but it 's not going to be . Cost/Len/Avg -37.967037200927734/80/-0.47458796501159667 INFO: SAMPLE 3: but it is expected to be expected to be held , but it 's not enough to be . Cost/Len/Avg -38.5172119140625/81/-0.47552113474151236 INFO: SAMPLE 4: but it is expected to be expected to be held , but it 's not going to be able . Cost/Len/Avg -40.010398864746094/85/-0.4707105748793658 INFO: SAMPLE 5: but it is expected to be expected to be held , but it will not be able to be . Cost/Len/Avg -40.27976989746094/79/-0.5098705050311511 INFO: SAMPLE 6: but it is expected to be expected to be held , but it will not be able to be able . Cost/Len/Avg -42.25888442993164/84/-0.5030819574991862 INFO: SAMPLE 7: but it is expected to be expected to be held , but it 's not going to be able to be . Cost/Len/Avg -43.747947692871094/91/-0.4807466779436384 INFO: SAMPLE 8: but it is expected to be expected to be held , but it 's not going to be able to go . Cost/Len/Avg -44.41551208496094/91/-0.48808255038418613 INFO: SAMPLE 9: but it is expected to be expected to be held , but it will not be able to be able to be . Cost/Len/Avg -46.169952392578125/90/-0.5129994710286458 INFO: SAMPLE 10: but it is expected to be expected to be held , but it 's not going to be able to be able . Cost/Len/Avg -46.30955505371094/96/-0.4823911984761556 INFO: SAMPLE 11: but it is expected to be expected to be held , but it will not be able to be able to be able . Cost/Len/Avg -48.63726806640625/95/-0.5119712428042763 INFO: SOURCE: 일부 펀@@ 드@@ 매@@ 니@@ 저는 " 기업 이@@ 윤@@ 이 증가@@ 해 주식@@ 시장이 활@@ 성@@ 화@@ 됐다 " 고 전했다 . INFO: TARGET: fund man@@ ag@@ ers said strong corporate profit growth was supporting stocks . INFO: SAMPLE 0: according to the company , the company said , " it 's a lot of growth . Cost/Len/Avg -37.93861770629883/82/-0.4626660695890101 INFO: SAMPLE 1: according to the company , the company said , " it 's a lot of recession . Cost/Len/Avg -37.9777946472168/85/-0.4467975840849035 INFO: SAMPLE 2: according to the company , the company said , " it 's a lot of economic growth . Cost/Len/Avg -38.13514709472656/91/-0.4190675504915007 INFO: SAMPLE 3: according to the company , the company said , " it 's a lot of interest rates . Cost/Len/Avg -38.815006256103516/90/-0.43127784729003904 INFO: SAMPLE 4: according to the company , the company said , " it 's a lot of recession . " Cost/Len/Avg -38.95130157470703/92/-0.4233837127685547 INFO: SAMPLE 5: according to the company , the company said , " it 's a lot of economic growth . " Cost/Len/Avg -39.129940032958984/98/-0.3992851023771325 INFO: SAMPLE 6: according to the company , the company said , " it 's a lot of interest rates . " Cost/Len/Avg -39.95598602294922/97/-0.41191738167988884 INFO: SAMPLE 7: according to the company , the company said , " it 's a lot of recession , " he said . Cost/Len/Avg -40.88528060913086/102/-0.4008360844032437 INFO: SAMPLE 8: according to the company , the company said , " it 's a lot of economic growth , " he said . Cost/Len/Avg -41.4897346496582/108/-0.38416420971905746 INFO: SAMPLE 9: according to the company , the company said , " it 's a lot of interest rates , " he said . Cost/Len/Avg -41.738468170166016/107/-0.3900791417772525 INFO: SAMPLE 10: according to the company , the company said , " it 's a lot of the economy , " he said . Cost/Len/Avg -42.56211471557617/104/-0.4092511030343863 INFO: SAMPLE 11: according to the company , the company said , " it 's a lot of economic growth , " the company said . Cost/Len/Avg -43.87964630126953/117/-0.37503971197666264 INFO: SOURCE: 이 바이러@@ 스는 서@@ 쪽으로 빠르게 퍼@@ 지고 있으며 , 관리들은 이것이 곧 서부 해안@@ 에@@ 까지 이를 것으로 우려하고 있다 . INFO: TARGET: the virus has been spreading quickly w@@ est@@ ward , with officials f@@ ear@@ ing it will soon reach the West Coast . INFO: SAMPLE 0: the virus is expected to be in the region . Cost/Len/Avg -21.322120666503906/44/-0.4845936515114524 INFO: SAMPLE 1: the virus is expected to be in the region , officials said . Cost/Len/Avg -23.85358238220215/61/-0.39104233413446143 INFO: SAMPLE 2: the virus is expected to be in the area , officials said . Cost/Len/Avg -24.124916076660156/59/-0.40889688265525687 INFO: SAMPLE 3: the virus is expected to be in the region , but officials say they are still in the region . Cost/Len/Avg -40.866912841796875/93/-0.4394291703419019 INFO: SAMPLE 4: the virus is expected to be in the region , but officials said it would be in the region . Cost/Len/Avg -41.33380126953125/91/-0.45421759636847525 INFO: SAMPLE 5: the virus is expected to be in the region , but officials say they are still in the area . Cost/Len/Avg -41.385223388671875/91/-0.45478267460078986 INFO: SAMPLE 6: the virus is expected to be in the region , but officials say they are still in the border . Cost/Len/Avg -41.601173400878906/93/-0.44732444517074094 INFO: SAMPLE 7: the virus is expected to be in the region , but officials said it would be in the area . Cost/Len/Avg -41.86735916137695/89/-0.470419765858168 INFO: SAMPLE 8: the virus is expected to be in the region , but officials say they are still in the town . Cost/Len/Avg -41.90886306762695/91/-0.4605369567871094 INFO: SAMPLE 9: the virus is expected to be in the region , but officials said it would be in the border . Cost/Len/Avg -41.95485305786133/91/-0.4610423412951794 INFO: SAMPLE 10: the virus is expected to be in the region , but officials say they are still in the city . Cost/Len/Avg -42.01584243774414/91/-0.4617125542609246 INFO: SAMPLE 11: the virus is expected to be in the region , but officials say they are still in the town of the region . Cost/Len/Avg -46.920806884765625/105/-0.44686482747395834 INFO: SOURCE: 잭슨 측의 한 변호사는 유@@ 효@@ 한 계약@@ 은 없@@ 었으며 돈@@ 은 대@@ 가@@ 성 없이 주@@ 어진 것이라고 말했습니다 . INFO: TARGET: a lawyer for Jackson says there was never a v@@ ali@@ d agreement , and that the money was given fre@@ ely . INFO: SAMPLE 0: the judge says it 's not clear that it 's not clear that it 's not enough . Cost/Len/Avg -35.635536193847656/91/-0.3915992988334907 INFO: SAMPLE 1: the judge says it 's not clear that it 's not clear that it 's not enough to be . Cost/Len/Avg -39.26071548461914/97/-0.40474964417133136 INFO: SAMPLE 2: the judge says it 's not clear that it was not clear that it 's not enough to be . Cost/Len/Avg -39.52196502685547/93/-0.4249673658801663 INFO: SAMPLE 3: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not enough . Cost/Len/Avg -45.02193832397461/117/-0.38480289165790266 INFO: SAMPLE 4: the judge says it 's not clear that it was not clear that it 's not clear that it 's not enough . Cost/Len/Avg -45.23143768310547/113/-0.40027820958500415 INFO: SAMPLE 5: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not a lot . Cost/Len/Avg -47.7430419921875/116/-0.41157794820851296 INFO: SAMPLE 6: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not enough to be . Cost/Len/Avg -48.963233947753906/123/-0.3980750727459667 INFO: SAMPLE 7: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not a little bit . Cost/Len/Avg -49.09718704223633/123/-0.3991641222946043 INFO: SAMPLE 8: the judge says it 's not clear that it was not clear that it 's not clear that it 's not enough to be . Cost/Len/Avg -49.15306091308594/119/-0.41305093204273896 INFO: SAMPLE 9: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not enough to be . " Cost/Len/Avg -52.79254150390625/130/-0.40609647310697117 INFO: SAMPLE 10: the judge says it 's not clear that it was not clear that it 's not clear that it 's not enough to be . " Cost/Len/Avg -53.00259017944336/126/-0.42065547761462985 INFO: SAMPLE 11: the judge says it 's not clear that it 's not clear that it 's not clear that it 's not a little bit . " Cost/Len/Avg -53.03470993041992/130/-0.40795930715707635 INFO: SOURCE: 총격@@ 범@@ 과 그의 여자@@ 친구 사이 의 말@@ 다@@ 툼 때문에 이번 공격이 발생@@ 했다는 , 확인@@ 되지 않은 보도가 있었다 . INFO: TARGET: un@@ confirmed reports say the attack followed an ar@@ gu@@ ment between the gunman and his girlfriend . INFO: SAMPLE 0: the incident was found in the attack . Cost/Len/Avg -16.111059188842773/39/-0.4131040817651993 INFO: SAMPLE 1: the incident was found in the attack , but it was not reported . Cost/Len/Avg -25.738344192504883/65/-0.39597452603853667 INFO: SAMPLE 2: the incident was found in the attack , but it was not immediately reported . Cost/Len/Avg -26.580915451049805/77/-0.34520669416947797 INFO: SAMPLE 3: the incident was found in the attack , but it was not immediately known . Cost/Len/Avg -27.21120262145996/74/-0.36771895434405355 INFO: SAMPLE 4: the incident was found in the attack , but the incident was not immediately known . Cost/Len/Avg -28.442596435546875/84/-0.3386023385184152 INFO: SAMPLE 5: the incident was found in the attack , but it was no evidence of the attack . Cost/Len/Avg -29.79374885559082/78/-0.3819711391742413 INFO: SAMPLE 6: the incident was found in the attack , but it was not immediately known as the attack . Cost/Len/Avg -32.138160705566406/88/-0.3652063716541637 INFO: SAMPLE 7: the incident was found in the attack , but it was not clear that the attack was not known . Cost/Len/Avg -34.64654541015625/92/-0.37659288489300274 INFO: SAMPLE 8: the incident was found in the attack , but it was not clear that the incident was not known . Cost/Len/Avg -34.85503005981445/94/-0.37079819212568566 INFO: SAMPLE 9: the incident was found in the attack , but it was not immediately clear that he was not known . Cost/Len/Avg -35.30630874633789/96/-0.3677740494410197 INFO: SAMPLE 10: the incident was found in the attack , but it was not immediately clear that the attack was not reported . Cost/Len/Avg -35.64750289916992/107/-0.33315423270252265 INFO: SAMPLE 11: the incident was found in the attack , but it was not clear that the attack was not immediately known . Cost/Len/Avg -35.765377044677734/104/-0.34389785619882435 INFO: SOURCE: 왕@@ 복@@ 선은 이틀@@ 간 우주 정거@@ 장을 조사한 뒤 호주 남@@ 동부 3@@ 36@@ km 위에서 작업을 종@@ 료@@ 할 예정이다 . INFO: TARGET: the shuttle 's two @-@ day ch@@ as@@ e of the space station ended about 2@@ 10 miles above southeastern Australia . INFO: SAMPLE 0: London , England ( CNN ) Prince Harry is expected to be in the first time after the first time of the space station after the first time of the space station . Cost/Len/Avg -52.85684585571289/160/-0.33035528659820557 /home/amitkumar.rs.cse17.iitbhu/nematus/nematus/train.py:290: RuntimeWarning: divide by zero encountered in true_divide i, sample, cost, len(sample), cost/len(sample)) INFO: SAMPLE 1: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 2: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 3: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 4: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 5: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 6: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 7: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 8: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 9: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 10: Cost/Len/Avg -10000000.0/0/-inf INFO: SAMPLE 11: Cost/Len/Avg -10000000.0/0/-inf INFO: SOURCE: 한국은 또한 전통적인 경제 논@@ 리에 명@@ 백@@ 히 거@@ 슬@@ 리는 이런 변화를 이@@ 룩@@ 해@@ 내@@ 길 바라고 있다 . INFO: TARGET: Korea also hopes to accompl@@ ish the shift in apparent de@@ fi@@ ance of con@@ v@@ enti@@ onal economic lo@@ gi@@ c . INFO: SAMPLE 0: South Korea also said it will also be a new economic crisis . Cost/Len/Avg -24.42115020751953/62/-0.3938895194761215 INFO: SAMPLE 1: South Korea also said that Korea is also likely to be a new economic crisis . Cost/Len/Avg -29.65618896484375/78/-0.3802075508313301 INFO: SAMPLE 2: South Korea also said that Korea is also likely to be a new economy . Cost/Len/Avg -29.85259246826172/70/-0.4264656066894531 INFO: SAMPLE 3: South Korea also said that Korea is also likely to make a new economic crisis . Cost/Len/Avg -29.87238121032715/80/-0.3734047651290894 INFO: SAMPLE 4: South Korea also said that Korea is also likely to be a major economic crisis . Cost/Len/Avg -29.930070877075195/80/-0.37412588596343993 INFO: SAMPLE 5: South Korea also said that Korea is also likely to take a new economic crisis . Cost/Len/Avg -30.075361251831055/80/-0.3759420156478882 INFO: SAMPLE 6: South Korea also said that Korea is also likely to get a new economic crisis . Cost/Len/Avg -30.59653091430664/79/-0.3872978596747676 INFO: SAMPLE 7: South Korea also said that Korea is also likely to have a new economic crisis . Cost/Len/Avg -30.841949462890625/80/-0.3855243682861328 INFO: SAMPLE 8: South Korea also said that Korea is also likely to be a new economic crisis in the world . Cost/Len/Avg -35.9841194152832/91/-0.3954298836844308 INFO: SAMPLE 9: South Korea also said that Korea is also likely to make a new economic crisis in the world . Cost/Len/Avg -36.09120178222656/93/-0.3880774385185652 INFO: SAMPLE 10: South Korea also said that Korea is also likely to be a new economic crisis in the country . Cost/Len/Avg -37.242855072021484/93/-0.40046080722603744 INFO: SAMPLE 11: South Korea also said that Korea is also likely to make a new economic crisis in the country . Cost/Len/Avg -37.32313919067383/95/-0.392875149375514 INFO: SOURCE: 크레@@ 디는 " 체포 당시 15@@ 달 된 아들을 안@@ 고 있었다 " 고 말했다 . INFO: TARGET: " I was holding my baby at the time , " said McC@@ ready of her 15 @-@ month @-@ old son . INFO: SAMPLE 0: he said he had been arrested last month . Cost/Len/Avg -18.11133575439453/42/-0.43122227986653644 INFO: SAMPLE 1: he said he had been arrested last year . Cost/Len/Avg -18.825273513793945/41/-0.45915301253155966 INFO: SAMPLE 2: he said he was taken to his wife . Cost/Len/Avg -19.07192039489746/35/-0.5449120112827845 INFO: SAMPLE 3: he said he had been arrested last month , " he said . Cost/Len/Avg -22.439355850219727/59/-0.3803280652579615 INFO: SAMPLE 4: he said he was taken to his wife , " he said . Cost/Len/Avg -22.90827751159668/52/-0.4405437982999362 INFO: SAMPLE 5: he said he had been arrested last month , " she said . Cost/Len/Avg -23.00881576538086/60/-0.38348026275634767 INFO: SAMPLE 6: he said he was taken to his wife , " she said . Cost/Len/Avg -23.43367576599121/53/-0.44214482577341907 INFO: SAMPLE 7: he said he had been arrested last month , " he said . " Cost/Len/Avg -28.204511642456055/66/-0.4273410854917584 INFO: SAMPLE 8: he said he was taken to his wife , " he said . " Cost/Len/Avg -28.393362045288086/59/-0.4812434244964082 INFO: SAMPLE 9: he said he had been arrested last month , " he said . " it was a child . Cost/Len/Avg -36.910091400146484/83/-0.44469989638730706 INFO: SAMPLE 10: he said he had been arrested last month , " he said . " it was a child . " Cost/Len/Avg -38.843902587890625/90/-0.43159891764322916 INFO: SAMPLE 11: he said he had been arrested last month , " he said . " it was not a child . Cost/Len/Avg -39.1064338684082/87/-0.44949923986676094 INFO: SOURCE: 이슬람 반군@@ 과의 이번 분쟁@@ 은 199@@ 0@@ 년 내@@ 전 이후 레바논@@ 에서 발생한 최악의 유혈@@ 충돌로 기록@@ 되고 있다 . INFO: TARGET: the sometimes @-@ fi@@ er@@ ce battles mark the worst internal violence since the end of Lebanon 's civil war in 1990 . INFO: SAMPLE 0: the Islamic militant group has been held in the northern city of southern Lebanon . Cost/Len/Avg -30.197078704833984/84/-0.35948903220040457 INFO: SAMPLE 1: the Islamic militant group has been held in the northern city of northern Lebanon . Cost/Len/Avg -30.308401107788086/84/-0.3608142989022391 INFO: SAMPLE 2: the Islamic militant group has been held in the northern city of northern Iraq . Cost/Len/Avg -30.530426025390625/81/-0.37691883981963736 INFO: SAMPLE 3: the Islamic militant group has been held in the northern city of northern Iraq in the region . Cost/Len/Avg -37.04274368286133/95/-0.38992361771432976 INFO: SAMPLE 4: the Islamic militant group has been held in the northern city of northern Lebanon , which has been in the region . Cost/Len/Avg -44.6102180480957/115/-0.3879149395486583 INFO: SAMPLE 5: the Islamic militant group has been held in the northern city of southern Lebanon , which has been in the region . Cost/Len/Avg -44.716522216796875/115/-0.38883932362432067 INFO: SAMPLE 6: the Islamic militant group has been held in the northern city of northern Iraq , which has been in the region . Cost/Len/Avg -44.74925994873047/112/-0.39954696382795063 INFO: SAMPLE 7: the Islamic militant group has been held in the northern city of northern Lebanon , which has been the worst in the region . Cost/Len/Avg -48.818878173828125/125/-0.390551025390625 INFO: SAMPLE 8: the Islamic militant group has been held in the northern city of southern Lebanon , which has been the worst in the region . Cost/Len/Avg -48.91313171386719/125/-0.3913050537109375 INFO: SAMPLE 9: the Islamic militant group has been held in the northern city of northern Iraq , which has been the worst in the region . Cost/Len/Avg -48.95859909057617/122/-0.4012999925457063 INFO: SAMPLE 10: the Islamic militant group has been held in the northern city of northern Lebanon , which has been the worst in the country . Cost/Len/Avg -49.13996124267578/126/-0.38999969240218874 INFO: SAMPLE 11: the Islamic militant group has been held in the northern city of southern Lebanon , which has been the worst in the country . Cost/Len/Avg -49.23438262939453/126/-0.3907490684872582 INFO: Seen 8 INFO: Seen 20 INFO: Seen 33 INFO: Seen 48 INFO: Seen 63 INFO: Seen 79 INFO: Seen 96 INFO: Seen 114 INFO: Seen 132 INFO: Seen 151 INFO: Seen 171 INFO: Seen 191 INFO: Seen 212 INFO: Seen 234 INFO: Seen 256 INFO: Seen 279 INFO: Seen 304 INFO: Seen 329 INFO: Seen 355 INFO: Seen 382 INFO: Seen 410 INFO: Seen 439 INFO: Seen 470 INFO: Seen 502 INFO: Seen 535 INFO: Seen 570 INFO: Seen 608 INFO: Seen 649 INFO: Seen 694 INFO: Seen 746 INFO: Seen 800 INFO: Seen 811 INFO: Seen 828 INFO: Seen 848 INFO: Seen 871 INFO: Seen 898 INFO: Seen 930 INFO: Seen 970 INFO: Seen 1000 INFO: Validation cross entropy (AVG/SUM/N_SENTS/N_TOKENS): 187.86940924072266 187869.40924072266 1000 33910 INFO: scripts/..//model/tmpv6367j94/model is not in all_model_checkpoint_paths. Manually adding it. INFO: Starting external validation. INFO: NOTE: Length of translations is capped to 200 INFO: Translated 40 sents INFO: Translated 80 sents INFO: Translated 120 sents INFO: Translated 160 sents INFO: Translated 200 sents INFO: Translated 240 sents INFO: Translated 280 sents INFO: Translated 320 sents INFO: Translated 360 sents INFO: Translated 400 sents INFO: Translated 440 sents INFO: Translated 480 sents INFO: Translated 520 sents INFO: Translated 560 sents INFO: Translated 600 sents INFO: Translated 640 sents INFO: Translated 680 sents INFO: Translated 720 sents INFO: Translated 760 sents INFO: Translated 800 sents INFO: Translated 840 sents INFO: Translated 880 sents INFO: Translated 920 sents INFO: Translated 960 sents INFO: Translated 1000 sents INFO: Translated 1000 sents in 23.049481868743896 sec. Speed 43.38492317070448 sents/sec Traceback (most recent call last): File "/home/amitkumar.rs.cse17.iitbhu/nematus/nematus/train.py", line 522, in