I want get graph_json from tvm.relay.from_tensorflow(), but I donnot install LLVM. What i can do?

mod, params = relay.frontend.from_tensorflow(graph_def, layout=layout, shape=shape_dict) target = ‘c’ target_host = ‘c’ with relay.build_config(opt_level=3): graph_json, lib, params = relay.build(mod, target=target, target_host=target_host, params=params)

but I get “TVMError: cannot convert type float32x512 to c type”, I only want to get graph_json.