GitLab

* add brier_score

* process brier_score

* brier score is working for N-sized class

* fxied brier score

* add TED to BigBench and Brier score to MMLU

* format

* Update metrics.py

* Update task.py

* Update generate_until_template_yaml

* Delete lm_eval/tasks/bigbench/aux_metric.py

* Update generate_until_template_yaml

* Update _default_template_yaml

[Refactor] Fix fewshot cot mmlu descriptions

[Refactor] Update docs ToC

[Refactor] Openai completions

Update README.md

[Refactor] add notebook for overview

Miscellaneous documentation updates

[refactor] mps requirement

[Refactor] Update README.md

We should probably revamp the distinction between "Basic" and "Advanced" usage, or maybe remove it all-together.

Removed `python -m ` since now calling `lm_eval` is enough.

fixes for sampler

[Refactor] Versioning

[Refactor] Urgent fix

[Refactor] Use correct HF model type for MBart-like models

[Refactor] Num_fewshot process

Rename bigbench.yml to default.yml

Changing the name since this is not actually the big bench format.