Viacheslav Kukushkin
Viacheslav Kukushkin
During mlcube development user builds & runs image with the same tag a lot of times, after every change. Right now only `mlcube/` folder hash is used for calculating temporary...
Solution for https://github.com/mlcommons/medperf/issues/496 : add `--valid` flag for entity listing: ``` medperf mlcube ls --valid medperf dataset ls --valid medperf benchmark ls --valid medperf result ls --valid ``` Waits for...
## Issue description I'm running medperf tutorials in WSL and face a strange behaviour when client and server start to fail randomly. As firstly I was thinking that's an internal...
### Describe the bug During [loss computations](https://github.com/mlcommons/GaNDLF/blob/master/GANDLF/compute/loss_and_metric.py#L113) it is assumed that loss function takes three params: prediction, target, params. However, it's not true for [CE](https://github.com/mlcommons/GaNDLF/blob/master/GANDLF/losses/regression.py#L59) loss that takes only prediction...
The PR adresses a few important bugs in how metrics are calculated. For now the fixes seem to work for binary classification tasks, but not guaranteed for regression/segmentation/multiclass classification (highly...
Currently (on 2024-05-24) the latest version of `docker` python package is 7.0.0. However, acouple of days ago new `requests` library was released, `2.32.0`. `docker` package [is not compatible](https://github.com/docker/docker-py/issues/3256) right now...
Adds a local web UI to use instead of CLI UI. On this stage, only read-only views are implemented: - list of benchmarks - list of mlcubes - list of...
Addresses https://github.com/mlcommons/medperf/issues/344 . Creates a `medperf mlcube edit` command that updates existing mlcube: - uid [required]. uid of desired mlcube - name [optional] - mlcube-file [optional] - parameters-file [optional] -...
`mlcube run` function is too long and complex right now, needs to be refactored (without behavior change)
Adds functionality for creating & managing datasets # Dataset submission - [x] "add new dataset" button on datasets list - [ ] "choose benchmark" step - [x] PoC (dropdown list)...