microsoft / rat-sql

A relation-aware semantic parsing model from English to SQL
https://arxiv.org/abs/1911.04942
MIT License
406 stars 117 forks source link

build Dockerfile causes segmentation fault #33

Closed tomerwolgithub closed 3 years ago

tomerwolgithub commented 3 years ago

Hi, while trying to run the Dockerfile using docker build -t ratsql . the process keeps failing due to a segmentation fault while trying to download Bert (step 7). Has anyone else encountered this issue?

Thanks!

Step 7/14 : RUN python -c "from transformers import BertModel; BertModel.from_pretrained('bert-large-uncased-whole-word-masking')"
 ---> Running in 43837f926306
To use data.metrics please install scikit-learn. See https://scikit-learn.org/stable/index.html
Segmentation fault (core dumped)
The command '/bin/sh -c python -c "from transformers import BertModel; BertModel.from_pretrained('bert-large-uncased-whole-word-masking')"' returned a non-zero code: 139
jayetri commented 3 years ago

@tomerwolgithub I am facing the same error but haven't come across a solution yet.

GhazalFallah commented 3 years ago

And me too.

mrlamm commented 3 years ago

+1

samhodge commented 3 years ago

it is a simple matter of updating to a later version of transformers than specified in the requirement.txt file.

papsipublic commented 3 years ago

Thanks for the tip. It worked with transformer 3.5.0, BUT NIOT with 3.4: 1) you need to specify in the setup.py : 'transformers == 3.5.0', 1.1) BUT the above line should be THE FIRST in the setup.opy in order to download the newest library, otherwise I got an incompatibility (and error) message: ERROR: transformers 3.5.0 has requirement sentencepiece==0.1.91, but you’ll have sentencepiece 0.1.94 which is incompatible.

Thanks again

samhodge commented 3 years ago

it also works to remove the version number from the transformer pip package, your solution is more complete, thanks for fleshing out the details

Sam