From ab09dc52f3329b916b3d3a292dce05cb99bac634 Mon Sep 17 00:00:00 2001 From: Benny Powers Date: Tue, 14 Jul 2026 19:55:36 +0300 Subject: [PATCH] =?UTF-8?q?bench:=20pass=20the=20resolved=20GLM=20engine?= =?UTF-8?q?=20path=20to=20eval=5Fglm.py=20=E2=80=94=20cwd-relative=20./glm?= =?UTF-8?q?=20default=20broke=20coli=20bench=20from=20other=20dirs=20(#203?= =?UTF-8?q?)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit cmd_bench builds the eval_glm.py command but did not pass --glm, so eval_glm.py fell back to ./glm which breaks when running from any directory other than the source tree. Pass the already-resolved GLM variable via --glm, matching --data and --snap. Assisted-By: Claude Opus 4.6 (1M context) --- c/coli | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/c/coli b/c/coli index 35dfd3c..8c89caf 100755 --- a/c/coli +++ b/c/coli @@ -585,7 +585,7 @@ def cmd_bench(a): print(f" {C.dim}downloading missing datasets: {', '.join(missing)}{C.r}") subprocess.call([py, os.path.join(TOOLS,"fetch_benchmarks.py"), "--out", a.data, "--tasks", ",".join(missing), "--limit", str(max(a.limit,200))]) - cmd=[py, os.path.join(TOOLS,"eval_glm.py"), "--snap",a.model, + cmd=[py, os.path.join(TOOLS,"eval_glm.py"), "--glm", GLM, "--snap",a.model, "--tasks", tasks, "--limit", str(a.limit), "--data", a.data] if a.ram: cmd+=["--ram",str(a.ram)] e=env_for(a)