crfm-helm 0.5.2__py3-none-any.whl → 0.5.3__py3-none-any.whl

helm/benchmark/static/schema_tables.yaml

@@ -99,47 +99,94 @@ metrics:
     display_name: METEOR
     short_display_name: METEOR
     description: METEOR
+    lower_is_better: false
   - name: f1
-    display_name: F1
-    short_display_name: F1
-    description: F1
+    display_name: BERTScore F1
+    short_display_name: BERTScore F1
+    description: BERTScore F1
+    lower_is_better: false
   - name: precision
     display_name: Precision
     short_display_name: Precision
     description: Precision
+    lower_is_better: false
   - name: recall
     display_name: Recall
     short_display_name: Recall
     description: Recall
+    lower_is_better: false
   - name: rouge1
     display_name: ROUGE-1
     short_display_name: ROUGE-1
     description: ROUGE-1
+    lower_is_better: false
   - name: rouge2
     display_name: ROUGE-2
     short_display_name: ROUGE-2
     description: ROUGE-2
+    lower_is_better: false
   - name: rougeL
     display_name: ROUGE-L
     short_display_name: ROUGE-L
     description: ROUGE-L
+    lower_is_better: false
   - name: rougeLsum
     display_name: ROUGE-Lsum
     short_display_name: ROUGE-Lsum
     description: ROUGE-Lsum
+    lower_is_better: false
   - name: bleu
     display_name: BLEU
     short_display_name: BLEU
     description: BLEU
+    lower_is_better: false
+  - name: accuracy
+    display_name: Accuracy
+    short_display_name: Accuracy
+    description: Accuracy
+    lower_is_better: false
+  - name: f1_macro
+    display_name: Macro F1
+    short_display_name: Macro F1
+    description: Macro F1
+    lower_is_better: false
+  - name: f1_micro
+    display_name: Micro F1
+    short_display_name: Micro F1
+    description: Micro F1
+    lower_is_better: false
+  - name: unsorted_list_exact_match
+    display_name: Unsorted List Exact Match
+    short_display_name: Exact Match
+    description: Unsorted List Exact Match
+    lower_is_better: false
+
+  # FinQA Accuracy
+  - name: program_accuracy
+    display_name: Program Accuracy
+    short_display_name: Program Accuracy
+    description: Program Accuracy
+    lower_is_better: false
+  - name: execution_accuracy
+    display_name: Execution Accuracy
+    short_display_name: Execution Accuracy
+    description: Execution Accuracy
+    lower_is_better: false
 
 perturbations: []
 
 metric_groups:
-  - name: accuracy
-    display_name: Accuracy
+  - name: main_metrics
+    display_name: Main Metrics
+    metrics:
+    - name: ${main_name}
+      split: __all__
+
+  - name: generation_metrics
+    display_name: Other Generation Metrics
     hide_win_rates: true
     metrics:
-    - name: meteor
+    - name: f1
       split: __all__
     - name: rouge1
       split: __all__
@@ -152,6 +199,17 @@ metric_groups:
     - name: bleu
       split: __all__
 
+  - name: classification_metrics
+    display_name: Classification Metrics
+    hide_win_rates: true
+    metrics:
+    - name: accuracy
+      split: __all__
+    - name: f1_macro
+      split: __all__
+    - name: f1_micro
+      split: __all__
+
   - name: efficiency
     display_name: Efficiency
     metrics:
@@ -180,13 +238,17 @@ run_groups:
     category: All Scenarios
     subgroups:
       - unitxt_cards.numeric_nlg
+      - unitxt_cards.tab_fact
+      - unitxt_cards.wikitq
+      - unitxt_cards.fin_qa
 
   - name: unitxt_cards.numeric_nlg
     display_name: NumericNLG
     short_display_name: NumericNLG
     description: "NumericNLG is a dataset for numerical table-to-text generation using pairs of a table and a paragraph of a table description with richer inference from scientific papers."
     metric_groups:
-      - accuracy
+      - main_metrics
+      - generation_metrics
       - efficiency
       - general_information
     environment:
@@ -198,3 +260,58 @@ run_groups:
       who: "?"
       when: "?"
       language: English
+
+  - name: unitxt_cards.tab_fact
+    display_name: TabFact
+    short_display_name: TabFact
+    description: "tab_fact is a large-scale dataset for the task of fact-checking on tables."
+    metric_groups:
+      - main_metrics
+      - classification_metrics
+      - efficiency
+      - general_information
+    environment:
+      main_name: accuracy
+      main_split: test
+    taxonomy:
+      task: "?"
+      what: "?"
+      who: "?"
+      when: "?"
+      language: English
+
+  - name: unitxt_cards.wikitq
+    display_name: WikiTableQuestions
+    short_display_name: WikiTableQuestions
+    description: "This WikiTableQuestions dataset is a large-scale dataset for the task of question answering on semi-structured tables."
+    metric_groups:
+      - main_metrics
+      - classification_metrics
+      - efficiency
+      - general_information
+    environment:
+      main_name: unsorted_list_exact_match
+      main_split: test
+    taxonomy:
+      task: "?"
+      what: "?"
+      who: "?"
+      when: "?"
+      language: English
+
+  - name: unitxt_cards.fin_qa
+    display_name: FinQA
+    description: The FinQA benchmark for numeric reasoning over financial data, with question answering pairs written by financial experts over financial reports [(Chen et al., 2021)](https://arxiv.org/abs/2109.00122/).
+    metric_groups:
+      - main_metrics
+      - efficiency
+      - general_information
+    environment:
+      main_name: program_accuracy
+      main_split: test
+    taxonomy:
+      task: question answering with numeric reasoning
+      what: financial reports
+      who: financial experts
+      when: 1999 to 2019
+      language: English

helm/benchmark/static/schema_thai.yaml

@@ -78,6 +78,7 @@ perturbations: []
 metric_groups:
   - name: accuracy
     display_name: Accuracy
+    hide_win_rates: true
     metrics:
       - name: ${main_name}
         split: ${main_split}
@@ -111,12 +112,32 @@ run_groups:
     description: Thai-language scenarios
     category: All scenarios
     subgroups:
+      - thai_exam
       - thai_exam_onet
       - thai_exam_ic
       - thai_exam_tgat
       - thai_exam_tpat1
       - thai_exam_a_level
 
+
+  - name: thai_exam
+    display_name: ThaiExam
+    description: >
+      Macro-averaged accuracy on all ThaiExam examinations.
+    metric_groups:
+      - accuracy
+      - efficiency
+      - general_information
+    environment:
+      main_name: exact_match
+      main_split: test
+    taxonomy:
+      task: question answering
+      what: n/a
+      who: n/a
+      when: "?"
+      language: Thai and English
+
   - name: thai_exam_onet
     display_name: ONET
     description: >