●Change8

b9715

📦 llama-cppView on GitHub →
✨ 3 features🔧 2 symbols

Summary

This release introduces CUDA support for the GGML col2im 1D operation and includes numerous pre-compiled binaries across multiple operating systems and hardware configurations.

✨ New Features

  • Added GGML_OP_COL2IM_1D operation support for CUDA.
  • Implemented optimization in CUDA col2im_1d using fast_div_modulo for index decomposition.
  • Tightened support checks, type matching, and destination contiguity for CUDA col2im_1d.

Affected Symbols

GGML_OP_COL2IM_1Dcol2im_1d