HPT-Large (with language) or HPT-Base (with language)

liruiw / HPT

Heterogeneous Pre-trained Transformer (HPT) as Scalable Policy Learner.

https://liruiw.github.io/hpt

MIT License

404 stars 21 forks source link

HPT-Large (with language) or HPT-Base (with language) #19

Closed AlbertYang0112 closed 1 day ago

AlbertYang0112 commented 4 days ago

The HPT-Large (with language) should be HPT-Base (with language) since the only difference between this model with HPT-Base is that this model has a language stem. The trunks are of the same size.

liruiw commented 3 days ago

good point! thanks for the catch!

liruiw commented 1 day ago

Updated in Readme