Irokobench: Benchmark Dataset for African languages (#2042)

* add afrixnli to task

* add chat completion

* remove chat completion -untested

* afrimmlu added

* afrimmlu folder update

* afrimmlu folder update

* updated prompt

* remove print

* add afrimgsm -direct

* add squad metric

* fix bash script

* remove direct util, update common yaml

* remove print

* add few show. metric fixes

* fix direct path, add bash script for gpt models

* added transate test

* update afrixnli tasks

* update afrixnli tasks

* update metrics for afrixnli

* prompt translations fix

* prompt translations fix

* filter and metric fix -mgsm

* remove squad metric

* remove squad metric

* add f1 score to mgsm

* add f1 score to mgsm

* update native-direct with lin

* change f1 function

* add lin to utils

* add utils

* remove test limit

* remove test configs

* add swahili to mmlu

* change eng to ewe in ewe yaml mmlu

* add squad metric to mgsm, remove whitespace filter

* added translate test

* added afrixnli_translate

* fix exact match valueError

* fix exact match valueError

* restructure mmlu folder

* spacing

* remove afrimmlu_translate folder

* add utility

* format task name, clean ups

* modefied mgsm

* update on afrimgsm

* update on afrimgsm

* removed utils

* other mgsm varieties

* other mgsm varieties

* adding trasnslate direct

* Update translate_direct_yaml

* add manual xnli prompt, add multichoice for openai models, and adapt multichoice metric for openai model

* edit for open models

* Update translate_direct_yaml

* add verbalizer for xnli

* change xnli from multiple choice to generate

* add manual accuracy scores

* revert xnli to multiple choice

* change afrimgsm utils

* revert xnli to multiple_choice

* cleanups and readmes

* remove openai fixes and unused regex

* pr review changes

* revert metrics.py, task.py and extraction.py to main version

---------
Co-authored-by: Israel Abebe Azime <azime@cg.uni-saarland.de>
Co-authored-by: Israel Abebe Azime <se.israel.abebe@gmail.com>

lm_eval/tasks/afrixnli/anli prompt/en-direct/afrixnli_en_direct_swa.yaml

+# Generated by utils.py
+dataset_name: swa
+include: afrixnli_en_direct_yaml
+task: afrixnli_en_direct_swa

lm_eval/tasks/afrixnli/anli prompt/en-direct/afrixnli_en_direct_twi.yaml

+# Generated by utils.py
+dataset_name: twi
+include: afrixnli_en_direct_yaml
+task: afrixnli_en_direct_twi

lm_eval/tasks/afrixnli/anli prompt/en-direct/afrixnli_en_direct_wol.yaml

+# Generated by utils.py
+dataset_name: wol
+include: afrixnli_en_direct_yaml
+task: afrixnli_en_direct_wol

lm_eval/tasks/afrixnli/anli prompt/en-direct/afrixnli_en_direct_xho.yaml

+# Generated by utils.py
+dataset_name: xho
+include: afrixnli_en_direct_yaml
+task: afrixnli_en_direct_xho

lm_eval/tasks/afrixnli/anli prompt/en-direct/afrixnli_en_direct_yaml

+group:
+    - xnli
+    - afrixnli
+    - afrixnli_en_direct
+dataset_path: masakhane/afrixnli
+dataset_name: null
+output_type: multiple_choice
+validation_split: validation
+test_split: test
+fewshot_split: validation
+doc_to_text: "{{premise}}\nQuestion: {{hypothesis}} True, False, or Neither?\nAnswer:"
+# True = entailment
+# False = contradiction
+# Neither = neutral
+doc_to_target: !function utils.doc_to_target
+doc_to_choice:
+  - "True"
+  - "Neither"
+  - "False"
+should_decontaminate: true
+doc_to_decontamination_query: premise
+metric_list:
+  - metric: f1
+    aggregation: !function utils.weighted_f1_score
+    average: weighted
+    higher_is_better: True
+    ignore_case: true
+    ignore_punctuation: true
+  - metric: acc
+    aggregation: mean
+    higher_is_better: true
+    ignore_case: true
+    ignore_punctuation: true
+metadata:
+  version: 1.0

lm_eval/tasks/afrixnli/anli prompt/en-direct/afrixnli_en_direct_yor.yaml

+# Generated by utils.py
+dataset_name: yor
+include: afrixnli_en_direct_yaml
+task: afrixnli_en_direct_yor

lm_eval/tasks/afrixnli/anli prompt/en-direct/afrixnli_en_direct_zul.yaml

+# Generated by utils.py
+dataset_name: zul
+include: afrixnli_en_direct_yaml
+task: afrixnli_en_direct_zul

lm_eval/tasks/afrixnli/anli prompt/en-direct/utils.py

+from sklearn.metrics import f1_score
+
+
+def doc_to_target(doc):
+    replacements = {
+        0: 'True',
+        1: 'Neither',
+        2: 'False'
+    }
+    return replacements[doc["label"]]
+
+
+def weighted_f1_score(items):
+    unzipped_list = list(zip(*items))
+    golds = unzipped_list[0]
+    preds = unzipped_list[1]
+    fscore = f1_score(golds, preds, average="weighted")
+    return fscore

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_amh.yaml

+# Generated by utils.py
+dataset_name: amh
+doc_to_choice: '{{[premise+", ትክክል? አዎ, "+hypothesis,premise+", ትክክል? እንዲሁም, "+hypothesis,premise+",
+  ትክክል? አይ, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_amh

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_eng.yaml

+# Generated by utils.py
+dataset_name: eng
+doc_to_choice: '{{[premise+", Right? Yes, "+hypothesis,premise+", Right? Also, "+hypothesis,premise+",
+  Right? No, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_eng

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_ewe.yaml

+# Generated by utils.py
+dataset_name: ewe
+doc_to_choice: '{{[premise+", Esɔ gbe? Ɛ̃, "+hypothesis,premise+", Esɔ gbe? Hã, "+hypothesis,premise+",
+  Esɔ gbe? Ao, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_ewe

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_fra.yaml

+# Generated by utils.py
+dataset_name: fra
+doc_to_choice: '{{[premise+", correct? Oui, "+hypothesis,premise+", correct? Aussi,
+  "+hypothesis,premise+", correct? Non, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_fra

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_hau.yaml

+# Generated by utils.py
+dataset_name: hau
+doc_to_choice: '{{[premise+", Daidai? Ee, "+hypothesis,premise+", Daidai? Haka kuma,
+  "+hypothesis,premise+", Daidai? A''a, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_hau

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_ibo.yaml

+# Generated by utils.py
+dataset_name: ibo
+doc_to_choice: '{{[premise+", Ziri ezi? Éè, "+hypothesis,premise+", Ziri ezi? Ọzọkwa,
+  "+hypothesis,premise+", Ziri ezi? Mba, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_ibo

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_kin.yaml

+# Generated by utils.py
+dataset_name: kin
+doc_to_choice: '{{[premise+", Nibyo? Yego, "+hypothesis,premise+", Nibyo? Na none,
+  "+hypothesis,premise+", Nibyo? Oya, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_kin

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_lin.yaml

+# Generated by utils.py
+dataset_name: lin
+doc_to_choice: '{{[premise+", Malamu? Iyo, "+hypothesis,premise+", Malamu? Lisusu,
+  "+hypothesis,premise+", Malamu? Te, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_lin

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_lug.yaml

+# Generated by utils.py
+dataset_name: lug
+doc_to_choice: '{{[premise+", Kituufu? Yee, "+hypothesis,premise+", Kituufu? N’ekirala,
+  "+hypothesis,premise+", Kituufu? Nedda, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_lug

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_orm.yaml

+# Generated by utils.py
+dataset_name: orm
+doc_to_choice: '{{[premise+", Sirrii? Eeyyee, "+hypothesis,premise+", Sirrii? Akkasumas,
+  "+hypothesis,premise+", Sirrii? Lakki, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_orm

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_sna.yaml

+# Generated by utils.py
+dataset_name: sna
+doc_to_choice: '{{[premise+", Chokwadi? Hongu, "+hypothesis,premise+", Chokwadi? Uye,
+  "+hypothesis,premise+", Chokwadi? Kwete, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_sna

lm_eval/tasks/afrixnli/anli prompt/native-direct/afrixnli_native_direct_sot.yaml

+# Generated by utils.py
+dataset_name: sot
+doc_to_choice: '{{[premise+", Nepile? E, "+hypothesis,premise+", Nepile? Hape, "+hypothesis,premise+",
+  Nepile? Tjhe, "+hypothesis]}}'
+include: afrixnli_native_direct_yaml
+task: afrixnli_native_direct_sot