Jiajia Qin

@vicary From the error message provided by @haoyunfeix , it's due to `wgpu` is behind of the latest WGSL spec. So maybe a better way is to report a bug...

@vladmandic I also can't reproduce it on my windows machine. @shurshilov I need your help to narrow down this issue to see which op's uniforms have the problem. Can you...

@vladmandic I see the first NaN op is `Minimum` in the log. I assume it's the first error happened. The second parameter of `Minimum` is a scalar. So I made...

@vladmandic @shurshilov Thanks for helping digging this. Now I have strong feeling that the NaN data is caused by [`Minimum`](https://github.com/tensorflow/tfjs/blob/master/tfjs-backend-webgl/src/kernels/Minimum.ts#L31)/[`Maximum`](https://github.com/tensorflow/tfjs/blob/master/tfjs-backend-webgl/src/kernels/Maximum.ts#L31)'s `isnan` check. In this [`CHECK_NAN_SNIPPET_PACKED`](https://github.com/tensorflow/tfjs/blob/master/tfjs-backend-webgl/src/binaryop_packed_gpu.ts#L24), you can see when `isNaN.xxx`...

@vladmandic I updated PR #6828. Two extra commits are uploaded. 1) [Remove NAN checking in binary](https://github.com/tensorflow/tfjs/pull/6828/commits/03ac08c29da4db39b1e87e528bc250633d01b3b2) 2) [Change the way of isNaN checking for Binary](https://github.com/tensorflow/tfjs/pull/6828/commits/8a6e23c094956a5727c9cd9a495788502f5b456e). Can you help to build...

@shurshilov Great! We have found the right place to fix this issue. Have a quick summary here if you are interested for the reason. 1. WEBGL_USE_SHAPES_UNIFORMS triggers this issue. But...

>btw, i dont really understand why is CHECK_NAN_SNIPPET used at all by default for packed binaries? its not used for unpacked binaries to start with. Both packed and unpacked binaries...

@vladmandic #6828 Have restored to method 2). Please take another try. I suppose it can be pass since `[commits 1 + 2 + 3: with patched kernel ops with new...

I just found a device which can reproduce this issue. I will give a debugging to see if I can have more findings. Please stay tuned.

@Linchenn Just tried webgpu (with #6760 ) on Intel TGL, about 700ms. Agree to not add it to DeepLabV3 in benchmarks for now. But there maybe some optimization opportunities. Can...