Minor error on example distillation script

[jaimeenahn]

Environment info

• transformers version: 4.5.1
• Platform: Linux-4.18.0-147.el8.x86_64-x86_64-with-Ubuntu-16.04-xenial
• Python version: 3.7.10
• PyTorch version (GPU?): 1.6.0 (True)
• Tensorflow version (GPU?): not installed (NA)
• Using GPU in script?: True
• Using distributed or parallel set-up in script?: True

Who can help

I think @VictorSanh might help since it's about a minor bug in distillation.

Information

Model I am using (Bert, XLNet ...):

The problem arises when using:

• the official example scripts: (give details below)
• my own modified scripts: (give details below)

examples/research_projects/distillation

The tasks I am working on is:

• [] an official GLUE/SQUaD task: (give the name)
• my own task or dataset: (give details below)

It's not GLUE/SQUaD but official BookCorpus and Wikipedia datasets from datasets

To reproduce

Steps to reproduce the behavior:

1. Convert concatenation of bookcorpus and Wikipedia text from datasets to txt file.
2. Separate it with \n
3. Run scripts following A. Preparing the data

Traceback (most recent call last):
  File "train.py", line 322, in <module>
    main()
  File "train.py", line 223, in main
    init_gpu_params(args)
  File "/volume/compression_and_distillation/transformers/examples/distillation/utils.py", line 55, in init_gpu_params
    if params.n_gpu <= 0:
AttributeError: 'Namespace' object has no attribute 'n_gpu'
Traceback (most recent call last):
  File "train.py", line 322, in <module>
    main()
  File "train.py", line 223, in main
    init_gpu_params(args)
  File "/volume/compression_and_distillation/transformers/examples/distillation/utils.py", line 55, in init_gpu_params
    if params.n_gpu <= 0:
AttributeError: 'Namespace' object has no attribute 'n_gpu'
Traceback (most recent call last):
  File "train.py", line 322, in <module>
    main()
  File "train.py", line 223, in main
    init_gpu_params(args)
  File "/volume/compression_and_distillation/transformers/examples/distillation/utils.py", line 55, in init_gpu_params
    if params.n_gpu <= 0:
AttributeError: 'Namespace' object has no attribute 'n_gpu'
Traceback (most recent call last):
  File "train.py", line 322, in <module>
    main()
  File "train.py", line 223, in main
    init_gpu_params(args)
  File "/volume/compression_and_distillation/transformers/examples/distillation/utils.py", line 55, in init_gpu_params
    if params.n_gpu <= 0:
AttributeError: 'Namespace' object has no attribute 'n_gpu'
Traceback (most recent call last):
  File "/usr/lib/python3.7/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/usr/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/venv/distill/lib/python3.7/site-packages/torch/distributed/launch.py", line 261, in <module>
    main()
  File "/home/venv/distill/lib/python3.7/site-packages/torch/distributed/launch.py", line 257, in main
    cmd=cmd)

Expected behavior

The error comes because of the inconsistency of variable name as n_gpu in the trainer.py script but gpus in parsing.

It can easily be solved changing gpus when parsing to n_gpu.