juney-nvidia
juney-nvidia
@pankajroark Hi, have you tried with the latest main branch or follow [this](https://github.com/NVIDIA/TensorRT-LLM/tree/main/examples/deepseek_v3#running-the-benchmark) guide to see whether the issue still exist? June
> Yes, in fact, these assertions are unnecessary. I will file a PR soon to fix it. Thanks, Chang! June
@WeiHaocheng @dc3671 @thorjohnsen Hi Fred/Zhenhuan/Thor Can you help review this PR from the community? Thanks June
@michaelfeil Thanks for submitting the MR. TRT-LLM has just become github firstly to make it easier for the community engagement. Can you help rebase your MR based on the latest...
@akhoroshev Hi, we plan to deprecate DS V1/V2 support, with only keeping the V3/R1 model support. So we may not accept this MR for now. Thanks June
@ming-wei Hi Ming, do you have any suggestion for this question? Thanks June
> QQ any benchmark compared with DeepGEMM on Hopper and Blackwell? Thanks. DeepGEMM only support Hopper WGMMA now, and on Blackwell we cannot directly use it. June
@nv-guomingz @tongyuantongyu to help review. cc @jiahanc for vis on this Hopper related effort.
> > DeepGEMM only support Hopper WGMMA now, and on Blackwell we cannot directly use it. > > Hi @juney-nvidia Thanks for the reply. Do you recommend to use DeepGEMM...