"this TensorFlow binary was not compiled to use" ... Aborted (core dumped)

[ruochongli]

Hi, I met this problem when I tried to run the following code on GPU:

python -m nmt.nmt \ --src=en --tgt=zh \ --vocab_prefix=nmt/wmt17/vocab.bpe.32000 \ --train_prefix=nmt/wmt17/train.tok.clean.bpe.32000 \ --dev_prefix=nmt/wmt17/newsdev2017.tok.bpe.32000 \ --test_prefix=/wmt17/newstest2017.tok.bpe.32000 \ --out_dir=nmt/nmt_model \ --num_train_steps=12000 \ --steps_per_stats=100 \ --num_layers=2 \ --num_units=128 \ --dropout=0.2 \ --metrics=bleu

Also, running the following code will give the same error:

python -m nmt.nmt \ --src=en --tgt=zh \ --hparams_path=nmt/standard_hparams/wmt16.json \ --out_dir=nmt/deen_gnmt \ --vocab_prefix=nmt/wmt17/vocab.bpe.32000 \ --train_prefix=nmt/wmt17/train.tok.clean.bpe.32000 \ --dev_prefix=nmt/wmt17/newsdev2017.tok.bpe.32000 \ --test_prefix=nmt/wmt17/newstest2017.tok.bpe.32000

The picture of the error:

Can anyone help? Is this a compiling problem? How can I tackle with it? Or is this a problem caused by my dataset?

Thank you very much.

[nashid]

I also had the same issue.