search

apache / tvm

Open deep learning compiler stack for cpu, gpu and specialized accelerators
https://tvm.apache.org/
Apache License 2.0
11.82k stars 3.48k forks source link

[Python][Relax] Update Rotary positional embedding scaling #17506

Closed tlopex closed 1 week ago

tlopex commented 3 weeks ago

This PR introduces two more styles of RoPE scaling: the gptj style and the yarn scale.

tlopex commented 3 weeks ago

cc @tqchen @MasterJH5574

tlopex commented 2 weeks ago

@MasterJH5574 Hi, Ruihang! Could you have a look at it and tell me how can I modify my code to pass the CI? Thanks!

tlopex commented 1 week ago

Thanks!@MasterJH5574 It works!