b9523
📦 llama-cppView on GitHub →
🐛 2 fixes🔧 3 symbols
Summary
This release refactors `hparams.n_layer`, removes the deprecated `n_layer_kv()` function in favor of `n_layer_all`, and includes various internal consistency fixes and bug corrections.
Migration Steps
- If using code that referenced `n_layer_kv()`, replace calls with `n_layer_all`.
🐛 Bug Fixes
- Fixed Step3.5 MTP in models.
- Fixed next layer count handling.