https://github.com/raynardj/blockbuster
Robust metric testing on model blocks
https://github.com/raynardj/blockbuster
Last synced: about 2 months ago
JSON representation
Robust metric testing on model blocks
- Host: GitHub
- URL: https://github.com/raynardj/blockbuster
- Owner: raynardj
- Created: 2026-04-11T06:03:21.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2026-04-11T06:51:08.000Z (2 months ago)
- Last Synced: 2026-04-11T08:23:21.743Z (2 months ago)
- Size: 7.81 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# BlockBuster 🎬
Robust metric testing on model blocks 🧱
We want to have a vanilla version of the model. 🍦
And we want to test and apply `hack`s to the vanilla model. 🍫
We produce a reports full of metrics on `hack`'s impact on the model.
* [Hacks](./designs/hacks.md) that we want to study.
* [Tasks](./designs/tasks.md) that we train and evaluate the model/optimizer on.
* [Metrics](./designs/metrics.md) we cared.
* [Vanillas](./designs/vanillas.md) what is concisdered baseline.
## Engineering
* Python [Packaging](./designs/packaging.md)
* Find price friendly way to run the experiments.[GPU Finding](./designs/gpu-finding.md)