b8508
📦 llama-cppView on GitHub →
🐛 2 fixes🔧 2 symbols
Summary
This release moves token embedding norms to the first layer in models and includes fixes for LLM_TENSOR_CONV1D and il indexing.
Migration Steps
- Models have been updated to move the token embedding norms to the first layer.
🐛 Bug Fixes
- Fixed LLM_TENSOR_CONV1D implementation.
- Fixed il indexing.