llama3_8b_chsnovel_q8_0_bs8_step120/config.json

{
    "model_type": "llama"
}