#!/bin/bash # set path and first lightx2v_path= model_path= export CUDA_VISIBLE_DEVICES=0 # set environment variables source ${lightx2v_path}/scripts/base/base.sh # Start API server with distributed inference service python -m lightx2v.api_server \ --model_cls wan2.1_distill \ --task i2v \ --model_path $model_path \ --config_json ${lightx2v_path}/configs/distill/wan_i2v_distill_4step_cfg.json \ --port 8000 \ --nproc_per_node 1 echo "Service stopped"