●Change8

v0.12.11

📦 ollama
✨ 6 features🐛 3 fixes🔧 6 symbols

Summary

Ollama 0.12.11 introduces logprobs support for API responses and adds opt-in Vulkan acceleration for expanded GPU compatibility.

Migration Steps

  1. To enable Vulkan support, set the environment variable OLLAMA_VULKAN=1 before running ollama serve.
  2. To use logprobs, add "logprobs": true to your API request payload.

✨ New Features

  • Support for log probabilities (logprobs) in Ollama and OpenAI-compatible APIs.
  • Support for top_logprobs to introspect alternative token likelihoods.
  • Opt-in Vulkan acceleration support for AMD, Intel, and iGPUs.
  • WebP image support in the new Ollama app.
  • Improved rendering performance in the app, specifically for code blocks.
  • GPU scheduling now prefers dedicated GPUs over integrated GPUs (iGPUs).

🐛 Bug Fixes

  • Fixed missing tool_call_id in the OpenAI-compatible API.
  • Fixed ollama create importing duplicate data from both consolidated.safetensors and other safetensor files.
  • The 'required' field in tool definitions is now omitted if not explicitly specified.

🔧 Affected Symbols

/api/generateOLLAMA_VULKANlogprobstop_logprobsollama createtool_call_id