如何使用Python部署自己的模型

[zyw11270106]

General Question

Hello，我微调asr模型后，部署自己微调后模型报错，一直未能解决，请帮忙指点下啊，感谢。 使用Python的进行调用，

调用如下 text = asr_executor( output_name, model='conformer_wenetspeech-zh-16k', config='/home/formal_ai/www/PaddleSpeech-develop/examples/aishell/asr1/conf/conformer1.yaml', ckpt_path='/home/formal_ai/www/PaddleSpeech-develop/examples/aishell/asr1/exp/conformer/checkpoints/5', device=paddle.get_device())

报错: [2022-09-20 15:15:59,625] [ INFO] - checking the audio file format...... [2022-09-20 15:15:59,628] [ INFO] - The sample rate is 16000 [2022-09-20 15:15:59,628] [ INFO] - The audio file format is right [2022-09-20 15:15:59,628] [ INFO] - start to init the model [2022-09-20 15:15:59,628] [ INFO] - /home/formal_ai/www/PaddleSpeech-develop/examples/aishell/asr1/conf/conformer.yaml [2022-09-20 15:15:59,629] [ INFO] - /home/formal_ai/www/PaddleSpeech-develop/examples/aishell/asr1/exp/conformer/checkpoints/5.pdparams [2022-09-20 15:15:59] [INFO] [_internal.py:224] 127.0.0.1 - - [20/Sep/2022 15:15:59] "POST /upload1 HTTP/1.1" 500 - Traceback (most recent call last): File "/home/anaconda3/envs/py37_pp/lib/python3.7/site-packages/flask/app.py", line 2091, in call return self.wsgi_app(environ, start_response) File "/home/anaconda3/envs/py37_pp/lib/python3.7/site-packages/flask/app.py", line 2076, in wsgi_app response = self.handle_exception(e) File "/home/anaconda3/envs/py37_pp/lib/python3.7/site-packages/flask/app.py", line 2073, in wsgi_app response = self.full_dispatch_request() File "/home/anaconda3/envs/py37_pp/lib/python3.7/site-packages/flask/app.py", line 1519, in full_dispatch_request rv = self.handle_user_exception(e) File "/home/anaconda3/envs/py37_pp/lib/python3.7/site-packages/flask/app.py", line 1517, in full_dispatch_request rv = self.dispatch_request() File "/home/anaconda3/envs/py37_pp/lib/python3.7/site-packages/flask/app.py", line 1503, in dispatch_request return self.ensure_sync(self.view_functions[rule.endpoint])(**req.view_args) File "/home/formal_ai/www/PaddleSpeech-develop/asr11.py", line 32, in upload_wav device=paddle.get_device()) File "/home/formal_ai/www/PaddleSpeech-develop/paddlespeech/cli/utils.py", line 346, in _warpper return executor_func(self, *args, **kwargs) File "/home/formal_ai/www/PaddleSpeech-develop/paddlespeech/cli/asr/infer.py", line 469, in call ckpt_path) File "/home/formal_ai/www/PaddleSpeech-develop/paddlespeech/cli/asr/infer.py", line 189, in _init_from_path self.config.decode.decoding_method = decode_method File "/home/anaconda3/envs/py37_pp/lib/python3.7/site-packages/yacs/config.py", line 141, in getattr raise AttributeError(name) AttributeError: decode

[zh794390558]

最新代码对应的配置文件见： https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/examples/aishell/asr1/conf/tuning/decode.yaml

用老的配置文件需要用老的代码，用新的代码的话，得用新的配置文件。

[zh794390558]

CLI里的调用的话，可以看下下载的模型里对应的配置。 打包好的模型类的model.yaml是通过./utils/generate_infer_yaml.py 这个生成的，和训练的是不一样的。需要注意下。

[zyw11270106]

最新代码对应的配置文件见： https://github.com/PaddlePaddle/PaddleSpeech/blob/develop/examples/aishell/asr1/conf/tuning/decode.yaml

用老的配置文件需要用老的代码，用新的代码的话，得用新的配置文件。

好的，我仔细检查一下配置文件，可能是在更换服务器时没有同步好

[zyw11270106]

CLI里的调用的话，可以看下下载的模型里对应的配置。 打包好的模型类的model.yaml是通过./utils/generate_infer_yaml.py 这个生成的，和训练的是不一样的。需要注意下。

我再检查一下配置文件，的确存在您说的这种问题，感谢。稍后有问题的话我再问您