v0.20.1
📦 ollamaView on GitHub →
✨ 1 features🐛 4 fixes🔧 2 symbols
Summary
This patch release introduces new benchmarking capabilities and resolves several parsing and build issues related to gemma4 and ROCm builds.
✨ New Features
- Benchmarking now includes prompt calibration, context size flag, and NumCtx reporting.
🐛 Bug Fixes
- Fixed gemma4 argument parsing when quoted strings contain the character ".
- Skipped cublasGemmBatchedEx during graph reservation in ggml.
- Fixed ROCm build for cublasGemmBatchedEx reserve wrapper.
- Reworked gemma4 tool call handling.