mrjoozy / llama32-vt

{ "num_predict": 256, "stop": [ "<|start_header_id|>", "<|end_header_id|>", "<|eot_id|>", "<|reserved_special_token" ] }