> The official project language is American English with ISO 8601 dates and metric units. Very good taste for an American.
NPU folks offen time say > it's not enough time to get new silicon designs specifically for <blahblah> Where blahblah stands for a model architecture that has caused a paradigm shift. When you need a new silicon for a…
OK, I am one of the developers in onnxruntime team. Perviously working on ROCm EP now has been transfered to QNN EP. The following is purely devrant and the opinions are mine. So ROCm already sucks whereas QNN sucks…
Yes.
This reminds me of Weight Agnostic Neural Networks https://weightagnostic.github.io/
Very unfriendly. The symbolic library (type of activations) requires a branching at the very core of the kernel. GPU will need to serialized on these operations warp-wise. To optimize, you might want to do a scan…
You run or your code runs, choose one and choose it wisely ;)
The memory is too small to be useful nowadays.
At the time of the dev of the EP, the tooling is not that good as current. I have imagined a pipeline that HLSL compiles down to DXIL then go through spirv cross and then target wide variety of mobile devices with…
Not exactly, both cl and glsl can be aot, but the runtime will be limited to some newer version and then the market coverage will be niche, those vendor are so lazy on updating the driver and fixing the compiler bugs...
circle lang is also very worth to check out.
Thoes "existing libraries" does not necessary mean stdc++, but some parallel primitive, and are essential to performance portability. For example, cub for scan and reduction, cutlass for dense linear algebra[1]. > I…
I am long sought after a CUDA or HIP compiler that target SPIR-V or DXIL. So that we can compile all thoes neural network kernels to almost all compute devices. The requirements are, 1. extend on C++17 this means…
Mixtral-8x7B support --> Support new model https://github.com/stanford-futuredata/megablocks/pull/45
Might be the training code related with the model https://github.com/mistralai/megablocks-public/tree/pstock/m...
You might also want to watch on helix, if PR 8675 is merged, then scheme will be the extension langauge, then we get the best of both world. Modal editing and scheme =) PR 8675:…
onnxruntime is written in C++. So it is not misleading =)
This papar is really interesting, especially so when you scroll to the Reference part ;)
I think you might be interesting in this interactive papar: https://willcrichton.net/nota/
Then we need a powerpoint virtual machine[1]. [1] https://www.andrew.cmu.edu/user/twildenh/PowerPointTM/Paper....
> The official project language is American English with ISO 8601 dates and metric units. Very good taste for an American.
NPU folks offen time say > it's not enough time to get new silicon designs specifically for <blahblah> Where blahblah stands for a model architecture that has caused a paradigm shift. When you need a new silicon for a…
OK, I am one of the developers in onnxruntime team. Perviously working on ROCm EP now has been transfered to QNN EP. The following is purely devrant and the opinions are mine. So ROCm already sucks whereas QNN sucks…
Yes.
This reminds me of Weight Agnostic Neural Networks https://weightagnostic.github.io/
Very unfriendly. The symbolic library (type of activations) requires a branching at the very core of the kernel. GPU will need to serialized on these operations warp-wise. To optimize, you might want to do a scan…
You run or your code runs, choose one and choose it wisely ;)
The memory is too small to be useful nowadays.
At the time of the dev of the EP, the tooling is not that good as current. I have imagined a pipeline that HLSL compiles down to DXIL then go through spirv cross and then target wide variety of mobile devices with…
Not exactly, both cl and glsl can be aot, but the runtime will be limited to some newer version and then the market coverage will be niche, those vendor are so lazy on updating the driver and fixing the compiler bugs...
circle lang is also very worth to check out.
Thoes "existing libraries" does not necessary mean stdc++, but some parallel primitive, and are essential to performance portability. For example, cub for scan and reduction, cutlass for dense linear algebra[1]. > I…
I am long sought after a CUDA or HIP compiler that target SPIR-V or DXIL. So that we can compile all thoes neural network kernels to almost all compute devices. The requirements are, 1. extend on C++17 this means…
Mixtral-8x7B support --> Support new model https://github.com/stanford-futuredata/megablocks/pull/45
Might be the training code related with the model https://github.com/mistralai/megablocks-public/tree/pstock/m...
You might also want to watch on helix, if PR 8675 is merged, then scheme will be the extension langauge, then we get the best of both world. Modal editing and scheme =) PR 8675:…
onnxruntime is written in C++. So it is not misleading =)
This papar is really interesting, especially so when you scroll to the Reference part ;)
I think you might be interesting in this interactive papar: https://willcrichton.net/nota/
Then we need a powerpoint virtual machine[1]. [1] https://www.andrew.cmu.edu/user/twildenh/PowerPointTM/Paper....