binh230/seallm-v3-4bit-qlora
Adapter vs SeaLLMs/SeaLLMs-v3-7B-Chat. binh230/seallm-v3-4bit-qlora is a LoRA adapter, not full weights: rank 16, on q_proj, v_proj in 28 layers (56 adapted matrices). It modifies SeaLLMs/SeaLLMs-v3-7B-Chat at load time.
Adapter vs SeaLLMs/SeaLLMs-v3-7B-Chat. binh230/seallm-v3-4bit-qlora is a LoRA adapter, not full weights: rank 16, on q_proj, v_proj in 28 layers (56 adapted matrices). It modifies SeaLLMs/SeaLLMs-v3-7B-Chat at load time.