Merge branch 'main' of https://github.com/EleutherAI/lm-evaluation-harness into t5v2-alt-plus

lm_eval/tasks/agieval/gaokao-mathcloze.yaml

+group:
+  - agieval
+  - agieval_cn
+task: agieval_gaokao_mathcloze
+dataset_path: hails/agieval-gaokao-mathcloze
+dataset_name: null
+output_type: generate_until
+training_split: null
+validation_split: null
+test_split: test
+doc_to_text: "{{query}}"
+doc_to_target: "{{answer}}"
+process_results: !function utils.process_results
+generation_kwargs:
+  max_gen_toks: 32
+  do_sample: False
+  temperature: 0.0
+  until:
+    - "Q:"
+metric_list:
+  - metric: acc
+    aggregation: mean
+    higher_is_better: true
+metadata:
+  version: 1.0

lm_eval/tasks/agieval/gaokao-mathqa.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_cn
+task: agieval_gaokao_mathqa
+dataset_path: hails/agieval-gaokao-mathqa

lm_eval/tasks/agieval/gaokao-physics.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_cn
+task: agieval_gaokao_physics
+dataset_path: hails/agieval-gaokao-physics

lm_eval/tasks/agieval/jec-qa-ca.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_cn
+task: agieval_jec_qa_ca
+dataset_path: hails/agieval-jec-qa-ca

lm_eval/tasks/agieval/jec-qa-kd.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_cn
+task: agieval_jec_qa_kd
+dataset_path: hails/agieval-jec-qa-kd

lm_eval/tasks/agieval/logiqa-en.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_nous
+  - agieval_en
+task: agieval_logiqa_en
+dataset_path: hails/agieval-logiqa-en

lm_eval/tasks/agieval/logiqa-zh.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_cn
+task: agieval_logiqa_zh
+dataset_path: hails/agieval-logiqa-zh

lm_eval/tasks/agieval/lsat-ar.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_nous
+  - agieval_en
+task: agieval_lsat_ar
+dataset_path: hails/agieval-lsat-ar

lm_eval/tasks/agieval/lsat-lr.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_nous
+  - agieval_en
+task: agieval_lsat_lr
+dataset_path: hails/agieval-lsat-lr

lm_eval/tasks/agieval/lsat-rc.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_nous
+  - agieval_en
+task: agieval_lsat_rc
+dataset_path: hails/agieval-lsat-rc

lm_eval/tasks/agieval/math.yaml

+group:
+  - agieval
+  - agieval_en
+task: agieval_math
+dataset_path: hails/agieval-math
+dataset_name: null
+output_type: generate_until
+training_split: null
+validation_split: null
+test_split: test
+doc_to_text: "{{query}}"
+doc_to_target: "{{answer}}"
+process_results: !function utils.process_results
+generation_kwargs:
+  max_gen_toks: 32
+  do_sample: False
+  temperature: 0.0
+  until:
+    - "Q:"
+metric_list:
+  - metric: acc
+    aggregation: mean
+    higher_is_better: true
+metadata:
+  version: 1.0

lm_eval/tasks/agieval/sat-en-without-passage.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_nous
+  - agieval_en
+task: agieval_sat_en_without_passage
+dataset_path: hails/agieval-sat-en-without-passage

lm_eval/tasks/agieval/sat-en.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_nous
+  - agieval_en
+task: agieval_sat_en
+dataset_path: hails/agieval-sat-en

lm_eval/tasks/agieval/sat-math.yaml

+include: aqua-rat.yaml
+group:
+  - agieval
+  - agieval_nous
+  - agieval_en
+task: agieval_sat_math
+dataset_path: hails/agieval-sat-math

lm_eval/tasks/agieval/utils.py

+# Answer parsing and normalization code, from
+# https://github.com/ruixiangcui/AGIEval/blob/main/src/
+# math_equivalence.py and post_process.py
+import re
+from typing import Dict, List
+
+import numpy as np
+
+
+def parse_math_answer(raw_string):
+    def remove_boxed(s):
+        left = "\\boxed{"
+        try:
+            assert s[: len(left)] == left
+            assert s[-1] == "}"
+            answer = s[len(left) : -1]
+            if "=" in answer:
+                answer = answer.split("=")[-1].lstrip(" ")
+            return answer
+        except Exception:
+            return None
+
+    def last_boxed_only_string(string):
+        idx = string.rfind("\\boxed")
+        if idx < 0:
+            idx = string.rfind("\\fbox")
+            if idx < 0:
+                return None
+        i = idx
+        right_brace_idx = None
+        num_left_braces_open = 0
+        while i < len(string):
+            if string[i] == "{":
+                num_left_braces_open += 1
+            if string[i] == "}":
+                num_left_braces_open -= 1
+                if num_left_braces_open == 0:
+                    right_brace_idx = i
+                    break
+            i += 1
+
+        if right_brace_idx is None:
+            retval = None
+        else:
+            retval = string[idx : right_brace_idx + 1]
+
+        return retval
+
+    def get_answer_with_dollar_sign(s):
+        first_pattern = "\$(.*)\$"
+        last_match = None
+        matches = re.findall(first_pattern, s)
+        if matches:
+            last_match = matches[-1]
+            if "=" in last_match:
+                last_match = last_match.split("=")[-1].lstrip(" ")
+        return last_match
+
+    def get_answer_without_dollar_sign(s):
+        last_match = None
+        if "=" in s:
+            last_match = s.split("=")[-1].lstrip(" ").rstrip(".")
+            if "\\n" in last_match:
+                last_match = last_match.split("\\n")[0]
+        else:
+            pattern = "(?:\\$)?\d+(?:\.\d+)?(?![\w\d])"
+            matches = re.findall(pattern, s)
+            if matches:
+                last_match = matches[-1]
+        return last_match
+
+    if "\\boxed" in raw_string:
+        answer = remove_boxed(last_boxed_only_string(raw_string))
+    else:
+        answer = get_answer_with_dollar_sign(raw_string)
+        if not answer:
+            answer = get_answer_without_dollar_sign(raw_string)
+    return answer
+
+
+# code from https://github.com/hendrycks/math/blob/main/modeling/math_equivalence.py
+def _fix_fracs(string):
+    substrs = string.split("\\frac")
+    new_str = substrs[0]
+    if len(substrs) > 1:
+        substrs = substrs[1:]
+        for substr in substrs:
+            new_str += "\\frac"
+            if substr[0] == "{":
+                new_str += substr
+            else:
+                try:
+                    assert len(substr) >= 2
+                except Exception:
+                    return string
+                a = substr[0]
+                b = substr[1]
+                if b != "{":
+                    if len(substr) > 2:
+                        post_substr = substr[2:]
+                        new_str += "{" + a + "}{" + b + "}" + post_substr
+                    else:
+                        new_str += "{" + a + "}{" + b + "}"
+                else:
+                    if len(substr) > 2:
+                        post_substr = substr[2:]
+                        new_str += "{" + a + "}" + b + post_substr
+                    else:
+                        new_str += "{" + a + "}" + b
+    string = new_str
+    return string
+
+
+def _fix_a_slash_b(string):
+    if len(string.split("/")) != 2:
+        return string
+    a = string.split("/")[0]
+    b = string.split("/")[1]
+    try:
+        a = int(a)
+        b = int(b)
+        assert string == "{}/{}".format(a, b)
+        new_string = "\\frac{" + str(a) + "}{" + str(b) + "}"
+        return new_string
+    except Exception:
+        return string
+
+
+def _remove_right_units(string):
+    # "\\text{ " only ever occurs (at least in the val set) when describing units
+    if "\\text{ " in string:
+        splits = string.split("\\text{ ")
+        assert len(splits) == 2
+        return splits[0]
+    else:
+        return string
+
+
+def _fix_sqrt(string):
+    if "\\sqrt" not in string:
+        return string
+    splits = string.split("\\sqrt")
+    new_string = splits[0]
+    for split in splits[1:]:
+        if split[0] != "{":
+            a = split[0]
+            new_substr = "\\sqrt{" + a + "}" + split[1:]
+        else:
+            new_substr = "\\sqrt" + split
+        new_string += new_substr
+    return new_string
+
+
+def _strip_string(string):
+    # linebreaks
+    string = string.replace("\n", "")
+    # print(string)
+
+    # remove inverse spaces
+    string = string.replace("\\!", "")
+    # print(string)
+
+    # replace \\ with \
+    string = string.replace("\\\\", "\\")
+    # print(string)
+
+    # replace tfrac and dfrac with frac
+    string = string.replace("tfrac", "frac")
+    string = string.replace("dfrac", "frac")
+    # print(string)
+
+    # remove \left and \right
+    string = string.replace("\\left", "")
+    string = string.replace("\\right", "")
+    # print(string)
+
+    # Remove circ (degrees)
+    string = string.replace("^{\\circ}", "")
+    string = string.replace("^\\circ", "")
+
+    # remove dollar signs
+    string = string.replace("\\$", "")
+
+    # remove units (on the right)
+    string = _remove_right_units(string)
+
+    # remove percentage
+    string = string.replace("\\%", "")
+    string = string.replace("\%", "")
+
+    # " 0." equivalent to " ." and "{0." equivalent to "{." Alternatively, add "0" if "." is the start of the string
+    string = string.replace(" .", " 0.")
+    string = string.replace("{.", "{0.")
+    # if empty, return empty string
+    if len(string) == 0:
+        return string
+    if string[0] == ".":
+        string = "0" + string
+
+    # to consider: get rid of e.g. "k = " or "q = " at beginning
+    if len(string.split("=")) == 2:
+        if len(string.split("=")[0]) <= 2:
+            string = string.split("=")[1]
+
+    # fix sqrt3 --> sqrt{3}
+    string = _fix_sqrt(string)
+
+    # remove spaces
+    string = string.replace(" ", "")
+
+    # \frac1b or \frac12 --> \frac{1}{b} and \frac{1}{2}, etc. Even works with \frac1{72} (but not \frac{72}1). Also does a/b --> \\frac{a}{b}
+    string = _fix_fracs(string)
+
+    # manually change 0.5 --> \frac{1}{2}
+    if string == "0.5":
+        string = "\\frac{1}{2}"
+
+    # NOTE: X/Y changed to \frac{X}{Y} in dataset, but in simple cases fix in case the model output is X/Y
+    string = _fix_a_slash_b(string)
+
+    return string
+
+
+def is_equiv(str1, str2, verbose=False):
+    if str1 is None and str2 is None:
+        print("WARNING: Both None")
+        return True
+    if str1 is None or str2 is None:
+        return False
+
+    str1, str2 = parse_math_answer(str1), parse_math_answer(str2)
+
+    try:
+        ss1 = _strip_string(str1)
+        ss2 = _strip_string(str2)
+        if verbose:
+            print(ss1, ss2)
+        return ss1 == ss2
+    except Exception:
+        return str1 == str2
+
+
+def process_results(doc: dict, results: List[str]) -> Dict[str, int]:
+    candidate = results[0]
+
+    gold = doc["answer"]
+
+    if not gold:
+        print(doc, candidate, gold)
+    if is_equiv(candidate, gold):
+        retval = 1
+    else:
+        retval = 0
+
+    results = {
+        "acc": retval,
+    }
+    return results
+
+
+# use a custom process_results() function, because AGIEval can have multiple valid answers
+def process_results_mcqa(doc, results):
+    results = [result[0] for result in results]
+
+    gold = doc["gold"]
+
+    acc = 1.0 if int(np.argmax(results)) in gold else 0.0
+    completion_len = np.array([float(len(i)) for i in doc["choices"]])
+    acc_norm = 1.0 if int(np.argmax(results / completion_len)) in gold else 0.0
+
+    return {
+        "acc": acc,
+        "acc_norm": acc_norm,
+    }

lm_eval/tasks/arithmetic/arithmetic_1dc.yaml

@@ -14,3 +14,5 @@ metric_list:
     higher_is_better: true
 metadata:
   version: 1.0
+dataset_kwargs:
+  trust_remote_code: true

lm_eval/tasks/arithmetic/arithmetic_2da.yaml

 include: arithmetic_1dc.yaml
 task: arithmetic_2da
 dataset_name: arithmetic_2da
+dataset_kwargs:
+  trust_remote_code: true

lm_eval/tasks/arithmetic/arithmetic_2dm.yaml

 include: arithmetic_1dc.yaml
 task: arithmetic_2dm
 dataset_name: arithmetic_2dm
+dataset_kwargs:
+  trust_remote_code: true

lm_eval/tasks/arithmetic/arithmetic_2ds.yaml

 include: arithmetic_1dc.yaml
 task: arithmetic_2ds
 dataset_name: arithmetic_2ds
+dataset_kwargs:
+  trust_remote_code: true

lm_eval/tasks/arithmetic/arithmetic_3da.yaml

 include: arithmetic_1dc.yaml
 task: arithmetic_3da
 dataset_name: arithmetic_3da
+dataset_kwargs:
+  trust_remote_code: true