pengxin99

@szha Thanks, and i will make appropriate modify.

@szhengac Thanks for review, > The loss need to be carefully rescaled, so that the final gradient is the average of the gradients w.r.t. tokens. and i want to know...

@szhengac @sxjscience could you please review the lasted commit if these changes work well, thanks~

@szha could you please help me found what is the problem or where did i go wrong? Thanks~ I use the new clip gradient ([pr #470](https://github.com/dmlc/gluon-nlp/pull/470)) for this multi GPUs...

@szha Thanks, and I will test GNMT performancem with multi GPUS , and then update this pr as soon as possible.

hi, @xhaihao , At first, we wanted to use` InsertPayloadToggle` in the struct `mfxExtMasteringDisplayColourVolume` to indicate whether the decoded clips have HDR SEI, but the ` InsertPayloadToggle` was designed for...

> Are RGB x coordinates stored in DisplayPrimariesX[3] in R=>G=>B order, or G=>B=>R order ? For the `DisplayPrimariesX[3]`, We just parse the SEI of clips storage information in order, and...

> is it possible to use a reserved byte in `mfxExtMasteringDisplayColourVolume` and `mfxExtContentLightLevelInfo` to indicate whether the SDK writes valid data back ? I think this way can be considered,...

@roywei thanks for you review, and I test our script on 8180-1s with 28 cores, below is the results: inference benchmark results： ``` INFO:root:lstm inference benchmark. INFO:root:For BS = 64,...

@kalyc @vandanavk @anirudhacharya I have addressed the review comments, could you please take a look? Thanks.