search

NVIDIA / TransformerEngine

A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper and Ada GPUs, to provide better performance with lower memory utilization in both training and inference.
https://docs.nvidia.com/deeplearning/transformer-engine/user-guide/index.html
Apache License 2.0
1.85k stars 309 forks source link

`inv_freq` of `RotaryPositionEmbedding` is hard-coded to 10k #849

Open shijie-wu opened 4 months ago

shijie-wu commented 4 months ago

theta in inv_freq of RotaryPositionEmbedding is hard-coded to 10k

https://github.com/NVIDIA/TransformerEngine/blob/50e7a3da8f3e04a054c9c7212bd80f71c6814a25/transformer_engine/pytorch/attention.py#L1371-L1377

ptrendx commented 4 months ago

@sudhakarsingh27 Could you take a look at it?