try lm-eval 0.4.11 to solve issue with transformers v5

dhuangnm · dtrifiro · commit b227674c27f0 · 2026-06-15T19:41:35.000+02:00
diff --git a/pyproject.toml b/pyproject.toml
@@ -6,7 +6,7 @@ authors = [
     {name = "Samuel Monson", email = "smonson@redhat.com"},
 ]
 dependencies = [
-    "lm-eval[api]==0.4.9.2",
+    "lm-eval[api]==0.4.11",
     "unitxt==1.22.0",
     "torch>=1.8",
 ]
@@ -15,7 +15,7 @@ readme = "README.md"
 
 [project.optional-dependencies]
 leaderboard = [
-    "lm-eval[ifeval,math,sentencepiece]>=0.4.8",
+    "lm-eval[ifeval,math,sentencepiece]>=0.4.11",
 ]
 
 # Use cpu pytorch to avoid downloading cuda

Original file line number	Diff line number	Diff line change
`@@ -6,7 +6,7 @@ authors = [`
`6`	`6`	`{name = "Samuel Monson", email = "smonson@redhat.com"},`
`7`	`7`	`]`
`8`	`8`	`dependencies = [`
`9`		`- "lm-eval[api]==0.4.9.2",`
	`9`	`+ "lm-eval[api]==0.4.11",`
`10`	`10`	`"unitxt==1.22.0",`
`11`	`11`	`"torch>=1.8",`
`12`	`12`	`]`
`@@ -15,7 +15,7 @@ readme = "README.md"`
`15`	`15`
`16`	`16`	`[project.optional-dependencies]`
`17`	`17`	`leaderboard = [`
`18`		`- "lm-eval[ifeval,math,sentencepiece]>=0.4.8",`
	`18`	`+ "lm-eval[ifeval,math,sentencepiece]>=0.4.11",`
`19`	`19`	`]`
`20`	`20`
`21`	`21`	`# Use cpu pytorch to avoid downloading cuda`