●Change8

b9915

📦 llama-cppView on GitHub →
✨ 1 features🔧 4 symbols

Summary

This release introduces CPU f16->f16 support for GGML_OP_SET_ROWS and refactors related compute functions. It also standardizes assertion usage within ggml by replacing assert() with GGML_ASSERT().

Migration Steps

  1. Replace usage of standard C assert() with GGML_ASSERT() in ggml code.

✨ New Features

  • Added support for CPU f16->f16 operations in GGML_OP_SET_ROWS.

Affected Symbols

GGML_OP_SET_ROWSggml_compute_forward_set_rows_f32()ggml_compute_forward_set_rows_f16()ggml_compute_forward_set_rows_impl()