v0.20.5-rc2
📦 ollamaView on GitHub →
✨ 3 features🐛 1 fixes🔧 2 symbols
Summary
This release introduces OpenClaw channel setup via `ollama launch openclaw` and enables flash attention and tool call repair for Gemma 4 models. A bug fix was also implemented for the `/save` command.
Migration Steps
- If using OpenClaw, utilize `ollama launch openclaw` for channel setup.
✨ New Features
- Introduced OpenClaw channel setup allowing connection to WhatsApp, Telegram, Discord, and other messaging channels via `ollama launch openclaw`.
- Enabled flash attention for Gemma 4 on compatible GPUs.
- Added tool call repair functionality for the Gemma 4 model family.
🐛 Bug Fixes
- Fixed the `/save `command when used with models based on safetensors architectures.