pyf98

/

librispeech_100_transducer_conformer

like 1

Hi, thanks for your nice work!
When using the pretrained model according to your guidance
./run.sh --skip_data_prep true --skip_train true --download_model pyf98/librispeech_100_transducer_conformer,
the following error arose. Do you have any idea about it?

# python3 -m espnet2.bin.asr_inference --batch_size 1 --ngpu 0 --data_path_and_name_and_type dump/raw/test_clean/wav.scp,speech,kaldi_ark --key_file exp/pyf98/librispeech_100_transducer_conformer/decode_asr_asr_model_valid.acc.ave/test_clean/logdir/keys.9.scp --asr_train_config exp/pyf98/librispeech_100_transducer_conformer/config.yaml --asr_model_file exp/pyf98/librispeech_100_transducer_conformer/valid.loss.ave_10best.pth --output_dir exp/pyf98/librispeech_100_transducer_conformer/decode_asr_asr_model_valid.acc.ave/test_clean/logdir/output.9 --config conf/decode_asr.yaml
# Started at Sat Oct 28 13:29:39 JST 2023
#
/home/smg/zengchang/apps/anaconda3/envs/espnet/bin/python3 /home/smg/zengchang/asr/espnet/espnet2/bin/asr_inference.py --batch_size 1 --ngpu 0 --data_path_and_name_and_type dump/raw/test_clean/wav.scp,speech,kaldi_ark --key_file exp/pyf98/librispeech_100_transducer_conformer/decode_asr_asr_model_valid.acc.ave/test_clean/logdir/keys.9.scp --asr_train_config exp/pyf98/librispeech_100_transducer_conformer/config.yaml --asr_model_file exp/pyf98/librispeech_100_transducer_conformer/valid.loss.ave_10best.pth --output_dir exp/pyf98/librispeech_100_transducer_conformer/decode_asr_asr_model_valid.acc.ave/test_clean/logdir/output.9 --config conf/decode_asr.yaml
2023-10-28 13:29:42,689 (asr:489) INFO: Vocabulary size: 5000
Traceback (most recent call last):
  File "/home/smg/zengchang/apps/anaconda3/envs/espnet/lib/python3.8/runpy.py", line 194, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/home/smg/zengchang/apps/anaconda3/envs/espnet/lib/python3.8/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/home/smg/zengchang/asr/espnet/espnet2/bin/asr_inference.py", line 902, in <module>
    main()
  File "/home/smg/zengchang/asr/espnet/espnet2/bin/asr_inference.py", line 898, in main
    inference(**kwargs)
  File "/home/smg/zengchang/asr/espnet/espnet2/bin/asr_inference.py", line 611, in inference
    speech2text = Speech2Text.from_pretrained(
  File "/home/smg/zengchang/asr/espnet/espnet2/bin/asr_inference.py", line 515, in from_pretrained
    return Speech2Text(**kwargs)
  File "/home/smg/zengchang/asr/espnet/espnet2/bin/asr_inference.py", line 178, in __init__
    beam_search_transducer = BeamSearchTransducer(
TypeError: type object argument after ** must be a mapping, not NoneType