mradermacher/sinhala-llama-2-7b-chat-hf-GGUF
Adapter vs Thimira/sinhala-llama-2-7b-chat-hf. Thimira/sinhala-llama-2-7b-chat-hf is a LoRA adapter, not full weights: rank 64, on q_proj, v_proj in 32 layers (64 adapted matrices). It modifies mradermacher/sinhala-llama-2-7b-chat-hf-GGUF at load time.