Issues with LM (TLG-rescoring)

[tonko22]

I'm trying to use CTC WFST-search for rescoring with compiled TLG graph using this tutorial: https://wenet-e2e.github.io/wenet/lm.html and passing these parameters to decoder:
config = { "model_path": f"wenet/{model_name}/final.zip", "dict_path": f"wenet/{model_name}/words.txt", "rescoring_weight": 1.0, "blank_skip_thresh": 0.98, "beam": 15.0, "lattice_beam": 7.5, "min_active": 10, "max_active": 7000, "ctc_weight": 0.5, "reverse_weight": 0.0, "chunk_size": -1, "fst_path": f"wenet/examples/aishell/s0/data/lang_test/TLG.fst" }

However I'm getting error:
`ERROR: FstImpl::ReadHeader: FST not of type vector, found qq: wenet/examples/aishell/s0/data/lang_test/TLG.fst
F1102 22:28:04.138978 26002 wenet_stt_lib.cpp:160] Check failed: fst != nullptr
*** Check failure stack trace: ***
@ 0x7f81d6cfb38d google::LogMessage::Fail()
@ 0x7f81d6cfd604 google::LogMessage::SendToLog()
@ 0x7f81d6cfaec0 google::LogMessage::Flush()
@ 0x7f81d6cfdd89 google::LogMessageFatal::~LogMessageFatal()
@ 0x7f81e83701b5 InitDecodeResourceFromSimpleJson()
@ 0x7f81e8380ebc WenetSTTModel::WenetSTTModel()
@ 0x7f81e83719bb wenet_stt__construct
@ 0x7f82021b7dec ffi_call_unix64
@ 0x7f82021b6f55 ffi_call
@ 0x7f82023d9e56 cdata_call
@ 0x5da58b _PyObject_FastCallKeywords
@ 0x54bc71 (unknown)
@ 0x552d2d _PyEval_EvalFrameDefault
@ 0x54cb89 _PyEval_EvalCodeWithName
@ 0x5dac6e _PyFunction_FastCallDict
@ 0x590713 (unknown)
@ 0x5da1c9 _PyObject_FastCallKeywords
@ 0x552fb7 _PyEval_EvalFrameDefault
@ 0x54c522 _PyEval_EvalCodeWithName
@ 0x54e933 PyEval_EvalCode
@ 0x6305a2 (unknown)
@ 0x630657 PyRun_FileExFlags
@ 0x6312cf PyRun_SimpleFileExFlags
@ 0x654232 (unknown)
@ 0x65458e _Py_UnixMain
@ 0x7f820422fb97 __libc_start_main
@ 0x5e0cca _start
@ (nil) (unknown)
Aborted

The same TLG-graph works fine when I'm using the default WeNet decoder.
Ubuntu 18.04.