Skip to content

Commit d8c6187

Browse files
author
semantic-release
committed
chore: release 0.81.2
1 parent d6e1b5b commit d8c6187

2 files changed

Lines changed: 21 additions & 1 deletion

File tree

CHANGELOG.md

Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,26 @@
11
# CHANGELOG
22

33

4+
## v0.81.2 (2026-03-29)
5+
6+
### Bug Fixes
7+
8+
- Batch_size must be multiple of num_generations, pad dataset if needed
9+
([#244](https://github.com/OpenAdaptAI/openadapt-evals/pull/244),
10+
[`d6e1b5b`](https://github.com/OpenAdaptAI/openadapt-evals/commit/d6e1b5bff59d672e5ec74126d35302f852ffe09a))
11+
12+
TRL requires generation_batch_size % num_generations == 0. With batch_size=1 and num_generations=4,
13+
TRL rejects it. Fix:
14+
15+
1. Set per_device_train_batch_size = num_generations (minimum valid) 2. Pad dataset by repeating
16+
tasks if len(dataset) < batch_size
17+
18+
With 1 task and num_generations=4: dataset padded to 4 rows, batch_size=4, generation_batch_size=4,
19+
4 % 4 == 0 ✓
20+
21+
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
22+
23+
424
## v0.81.1 (2026-03-29)
525

626
### Bug Fixes

pyproject.toml

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
44

55
[project]
66
name = "openadapt-evals"
7-
version = "0.81.1"
7+
version = "0.81.2"
88
description = "Evaluation infrastructure for GUI agent benchmarks"
99
readme = "README.md"
1010
requires-python = ">=3.10"

0 commit comments

Comments
 (0)