Change8

b9123

📦 llama-cppView on GitHub →
2 features🐛 1 fixes🔧 2 symbols

Summary

This release enables running the gpt-oss-20b model via ggml-webgpu and includes various pre-compiled binaries for numerous operating systems and hardware configurations.

✨ New Features

  • Enabled running gpt-oss-20b via ggml-webgpu.
  • Refactored mulmat-q operation in ggml-webgpu.

🐛 Bug Fixes

  • Disabled test-backend-ops on ubuntu-24-webgpu.

Affected Symbols