NVIDIA / Megatron-LM

Ongoing research training transformer models at scale
https://docs.nvidia.com/megatron-core/developer-guide/latest/user-guide/index.html#quick-start
Other
9.23k stars 2.08k forks source link

[QUESTION] Using segformer segmentation models #868

Open cporrasn opened 2 weeks ago

cporrasn commented 2 weeks ago

Your question It is possible to use the segmentation models that exist in megatron, I have found a main with the refinement, but I am not clear how a simple inference can be proven. Do these segmentation models use tensor parallelism?