local variable 'tag' in asr_engine.py

lang101 commented 2 years ago

Error occured with 'cfg_path', 'am_model', 'am_params' are not none. So, local variable 'tag' must be assigned.

zh794390558 commented 2 years ago

please give more info.

lang101 commented 2 years ago

https://github.com/PaddlePaddle/PaddleSpeech/blob/44ee5cd80596b93e8ce3e3f858acba509dfe21e9/paddlespeech/server/engine/asr/online/asr_engine.py#L75-L77 These parameters (cfg_path, am_model, am_params) not empty (configured in application.yaml) occured error in L113-L114

lm_url = pretrained_models[tag]['lm_url']
lm_md5 = pretrained_models[tag]['lm_md5']

Same as https://github.com/PaddlePaddle/PaddleSpeech/blob/44ee5cd80596b93e8ce3e3f858acba509dfe21e9/paddlespeech/server/engine/asr/paddleinference/asr_engine.py#L112-L113

https://github.com/PaddlePaddle/PaddleSpeech/blob/44ee5cd80596b93e8ce3e3f858acba509dfe21e9/paddlespeech/cli/asr/infer.py#L258-L259

WilliamZhang06 commented 2 years ago

https://github.com/PaddlePaddle/PaddleSpeech/blob/44ee5cd80596b93e8ce3e3f858acba509dfe21e9/paddlespeech/server/engine/asr/online/asr_engine.py#L75-L77

These parameters (cfg_path, am_model, am_params) not empty (configured in application.yaml) occured error in L113-L114
lm_url = pretrained_models[tag]['lm_url']
lm_md5 = pretrained_models[tag]['lm_md5']
Same as

https://github.com/PaddlePaddle/PaddleSpeech/blob/44ee5cd80596b93e8ce3e3f858acba509dfe21e9/paddlespeech/server/engine/asr/paddleinference/asr_engine.py#L112-L113

https://github.com/PaddlePaddle/PaddleSpeech/blob/44ee5cd80596b93e8ce3e3f858acba509dfe21e9/paddlespeech/cli/asr/infer.py#L258-L259

Please provide a screenshot of the Python error. need to reproduce this problem

lang101 commented 2 years ago

capplication.yaml

asr_online:
    model_type: 'deepspeech2online_aishell'
    am_model: '/data/model/deepspeech2online_aishell-zh-16k/exp/deepspeech2_online/checkpoints/avg_1.jit.pdmodel' # the pdmodel file of am static model [optional]
    am_params: '/data/model/deepspeech2online_aishell-zh-16k/exp/deepspeech2_online/checkpoints/avg_1.jit.pdiparams'  # the pdiparams file of am static model [optional]
    lang: 'zh'
    sample_rate: 16000
    cfg_path: '/data/model/deepspeech2online_aishell-zh-16k/model.yaml'
    decode_method:
    force_yes: True

    am_predictor_conf:
        device: 'cpu'  # set 'gpu:id' or 'cpu'
        switch_ir_optim: True
        glog_info: False  # True -> print glog
        summary: True  # False -> do not show predictor config

results

(test) root@0412fc2dad04:/data/test/PaddleSpeech# paddlespeech_server start --config_file ./paddlespeech/server/conf/application.yaml

[nltk_data] Error loading averaged_perceptron_tagger: <urlopen error [nltk_data] [Errno 101] Network is unreachable> [nltk_data] Error loading cmudict: <urlopen error [Errno 101] Network [nltk_data] is unreachable> Traceback (most recent call last): File "/usr/local/miniconda3/envs/test/bin/paddlespeech_server", line 33, in sys.exit(load_entry_point('paddlespeech', 'console_scripts', 'paddlespeech_server')()) File "/data/test/PaddleSpeech/paddlespeech/server/entry.py", line 36, in server_execute status = 0 if com['_entry']().execute(sys.argv[idx:]) else 1 File "/data/test/PaddleSpeech/paddlespeech/server/bin/paddlespeech_server.py", line 84, in execute if self.init(config): File "/data/test/PaddleSpeech/paddlespeech/server/bin/paddlespeech_server.py", line 75, in init if not init_engine_pool(config): File "/data/test/PaddleSpeech/paddlespeech/server/engine/engine_pool.py", line 37, in init_engine_pool if not ENGINE_POOL[engine].init(config=config[engine_and_type]): File "/data/test/PaddleSpeech/paddlespeech/server/engine/asr/online/asr_engine.py", line 291, in init self.executor._init_from_path( File "/data/test/PaddleSpeech/paddlespeech/server/engine/asr/online/asr_engine.py", line 113, in _init_from_path lm_url = pretrained_models[tag]['lm_url'] UnboundLocalError: local variable 'tag' referenced before assignment

ws_capplication.yaml

asr_online:
    model_type: 'deepspeech2online_aishell'
    am_model: '/data/model/deepspeech2online_aishell-zh-16k/exp/deepspeech2_online/checkpoints/avg_1.jit.pdmodel'     # the pdmodel file of am static model [optional]
    am_params: '/data/model/deepspeech2online_aishell-zh-16k/exp/deepspeech2_online/checkpoints/avg_1.jit.pdiparams'  # the pdiparams file of am static model [optional]
    lang: 'zh'
    sample_rate: 16000
    cfg_path: '/data/model/deepspeech2online_aishell-zh-16k/model.yaml'
    decode_method:
    force_yes: True

    am_predictor_conf:
        device: 'cpu' # set 'gpu:id' or 'cpu'
        switch_ir_optim: True
        glog_info: False  # True -> print glog
        summary: True  # False -> do not show predictor config

    chunk_buffer_conf:
        frame_duration_ms: 80
        shift_ms: 40
        sample_rate: 16000
        sample_width: 2

    vad_conf:
        aggressiveness: 2
        sample_rate: 16000
        frame_duration_ms: 20
        sample_width: 2
        padding_ms: 200
        padding_ratio: 0.9

results

(test) root@0412fc2dad04:/data/test/PaddleSpeech# paddlespeech_server start --config_file ./paddlespeech/server/conf/ws_application.yaml

Same as previously shown

lym0302 commented 2 years ago

The problem of tag has been solved in the latest code, thank you for raising the question

lym0302 commented 2 years ago

The problem of tag has been solved in the latest code. 'tag' should be written above 'if',fixed.

PaddlePaddle / PaddleSpeech