How to using CPU for elmo?

[ybdesire]

I run below cmd at my CPU environment as the readme said

python -m preprocess.get_elmo_vec data/conll2003/

but get errors below, since I do not have GPU

Traceback (most recent call last):
  File "/root/anaconda3/envs/pt_lstmcrf/lib/python3.7/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/root/anaconda3/envs/pt_lstmcrf/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/tmp/pytorch_lstmcrf/preprocess/get_elmo_vec.py", line 99, in <module>
    get_vector()
  File "/home/tmp/pytorch_lstmcrf/preprocess/get_elmo_vec.py", line 75, in get_vector
    elmo = load_elmo(cuda_device)
  File "/home/tmp/pytorch_lstmcrf/preprocess/get_elmo_vec.py", line 38, in load_elmo
    return ElmoEmbedder(cuda_device=cuda_device)
AssertionError:
Found no NVIDIA driver on your system. Please check that you
have an NVIDIA GPU and installed a driver from
http://www.nvidia.com/Download/index.aspx

Would you share the steps to get elmo word embedding by CPU ?

[allanj]

You can simply change cuda_device to -1 in https://github.com/allanj/pytorch_lstmcrf/blob/master/preprocess/get_elmo_vec.py#L74

By replace

cuda_device = 0

with

cuda_device = -1

Such that it will be using CPU.
Feel free to reopen the issue if you run into other problems.

[ybdesire]

Thanks @allanj for the help.