d2l-tvm

d2l-tvm copied to clipboard

The measured GFLOPS is larger than the peak performance

1

In section http://tvm.d2l.ai/chapter_cpu_schedules/arch.html#performance, the measured GFLOPS is 774.949674567942 and the peak performance is 640. Why is the measured GFLOPS larger than the peak performance? And it says "As can be...

lee-bin

Are there any plan to update?

1

I have read all contents updated right now and benefit from it a lot. As a result i wonder if there is a plan to finish the rest of chapters?

leadalion

the-mountain-rcj

Colab pip install issue

2

When pip installing on colab ``` !pip install https://tvm-repo.s3-us-west-2.amazonaws.com/tvm-0.7.dev1-cp37-cp37m-linux_x86_64.whl https://tvm-repo.s3-us-west-2.amazonaws.com/topi-0.7.dev1-py3-none-any.whl ``` it will show this error ``` ERROR: tvm-0.7.dev1-cp37-cp37m-linux_x86_64.whl is not a supported wheel on this platform. ``` I remember...

Zha0q1

How To: Dynamic Batch, Low Bit Quantization Calibration, FP16 Inference, Training via TVM

3

Most wanted: Dynamic Batch Low Bit Quantization Calibration FP16 Inference Training via TVM Low level: Adapt to a new CPU/GPU Architecture, such as MIPS, Specialization RISC-V(with special SIMD implementation).

kalcohol