1.10.1 1.10.0
camel-ai/owlSep 7, 2026 → Sep 4, 20260 days between releases
Generating AI comparison
Analyzing what changed between 1.10.1 and 1.10.0. This page updates automatically — usually within a minute.
The Two Releases
v1.10.1: CUDA Compile Fixes & Memory Reduction
Sep 7, 2026· 1 month ago
A patch release that fixes CUDA compilation issues, enabling `torch.compile` on CUDA with the default `multi_scale=True` recipe. It also reduces peak CUDA memory in segmentation loss and fixes device-sync and strategy-selection issues for the experimental TPU/XLA training path.
Full release notesv1.10.0: Faster Training & Inference
Sep 4, 2026· 2 months ago
A training-focused release that significantly speeds up training and inference, reducing training time by approximately 44% and inference speed by up to 46% for large frames. It introduces a GPU batched linear-assignment solver and changes default training parameters.
Full release notes