Text Generation
Transformers
Safetensors
English
olmo
Inference Endpoints

Intermediate checkpoints for HF model

#1
by przvl - opened

Thanks to the whole team for the great work on the OLMo models!

On the model card you state:

We are releasing many checkpoints for these models, for every 1000 training steps.
These have not yet been converted into Hugging Face Transformers format, but are available in allenai/OLMo-7B.

Are you still converting the checkpoints to HF format? Would be really helpful for easily comparing different checkpoints with transformers (also for the 1B model).

+1 would like to follow up on this as I would like to use the HF format models!

Allen Institute for AI org

As of today, we have released almost all the checkpoints of the newer allenai/OLMo-1.7-7B-hf model. The original 1B model will probably be next.

If you have any particular intermediate checkpoints you are interested in using, then one option is to convert these to HF format yourself (it takes maybe 5-10 mins per checkpoint). The instructions are in Checkpoints.md. The idea is to find the official checkpoint you want in https://github.com/allenai/OLMo/blob/main/checkpoints/official and then use convert_olmo_to_hf_new.py to convert it to HF format.

Sign up or log in to comment