ck
ck copied to clipboard
mlcommons
Collective Mind (CM) is a small, modular, cross-platform and decentralized workflow automation framework with a human-friendly interface and reusable automation recipes to make it easier to build, run...
When using `Python3.9`, the minimum `PyTorch==1.7.1` version, but in the config in the `CK / local / venv / mlperf-inference / CK / ck-mlops / package / lib-python-torch / .cm...
I followed the instructions https://github.com/mlcommons/ck/blob/master/docs/mlperf-automation/reproduce/ck-object-detection-rpi4-tflite.md for Raspian OS Pi 4 and I came across these errors. One was when installing `ck install package --tags=lib,mlperf,loadgen,static` it produced an error on the...
As discussed during our conf call, to pick the corresponding variation from the install script, we can create empty variations inside get scripts for all the variations in the corresponding...
As per convention we are mostly using only env variables to pass in information to any CM script and these variables are in capital letters. When the number of inputs...
Preparing agenda for the weekly meetings of our [MLCommons workgroup](https://github.com/mlcommons/ck/blob/master/docs/mlperf-education-workgroup.md) to modularize MLPerf benchmarks, make it easier to plug in any ML task, model, data set, framework, compiler and hardware...
The goal of this project (summer internship of Thomas Zhu from Oxford University) is to prototype CM-based modularization of the MLPerf inference benchmark to connect [C++ loadgen](https://github.com/mlcommons/inference/tree/master/loadgen) with different ML...
As discussed during the conf-calls, we should ass test.sh files inside individual script folders which can test the basic script functionality. Need to think how to handle variations and versions...
Suppose a script is having dependency on onnxruntime and pytorch and system default python is 3.10 - onnxruntime requires python
I suggest to create a catalog of [CM automations](https://github.com/mlcommons/ck/tree/master/cm-mlops/script) for MLPerf/ML/AI artifacts and tools that can be reused across projects. We can add it to the root README with links...
I plan to add an unoptimized reference implementation of image classification with TVM to our CK2(CM)-based MLPerf inference benchmark automation.