Musashi Hinck

icon company Intel Labs icon location AI Research Scientist at Intel Labs | Prev: Postdoc at Princeton University | DPhil Oxford

Results 4 comments of Musashi Hinck

reached the type-length limit while instantiating `tokio::runtime::Handle::enter::<...sult::Result<(), anyhow::Error>>`

Also getting this error, both on Arch Linux armv7 and Fedora 33.

error when adding geom_signif to a plot

Running into the same issue. Pretty sure it's related to #63, wasn't able to make that solution work for me though.

Showing raw OpenAI requests / responses

I would be happy to take this up if you're looking for an extra pair of hands!

How to load an OPT-IML checkpoint and use it for inference?

Using the `convert_to_singleton.py` script above runs into a different error when used directly on the IML checkpoints: ``` ['--model-parallel-size', '2', '--distributed-world-size', '2', '--ddp-backend', 'pytorch_ddp', '--task', 'language_modeling', '--bpe-merges', '../OPT-IML-30B/gpt2-merges.txt', '--merges-filename', '../OPT-IML-30B/gpt2-merges.txt',...