Transformers: How to use gpt-2-xl with run_generation.py

[avianion]

❓ Questions & Help

I am currently running the run_generation.py and putting the model as gpt-2, however I am not satisfied with the results frm the small model.

How can I insert the big model instead?

[LysandreJik]

You're probably using the command:

python run_generation.py --model_type=gpt2 --model_name_or_path=gpt2

to use GPT2-XL for generation you would change the last argument to gpt2-xl:

python run_generation.py --model_type=gpt2 --model_name_or_path=gpt2-xl

[avianion]

ah. its because the latest version isnt available on pypi yet. and i pip installed it. 2.1.1 is the version I've got installed.

Is there any way to force pip to install the latest version that was commited 21 hours ago.

[LysandreJik]

Ah yes we're yet to release a pypi version with GPT2-XL. To install from source you would have to do:

pip install git+https://github.com/huggingface/transformers

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.